Spectrally efficient high capacity wireless communication systems

Background of the Invention

0001This invention relates to wireless communication systems and more particularly, to using antenna arrays and signal processing to dramatically increase the capacity and performance of wireless communication systems.

0002Wireless communication systems can be used to complement and in some instances replace conventional wired communication systems in areas where conventional wire-line systems are unavailable, unreliable, or excessively expensive. Examples of such areas are: rural areas with a small number of widespread users, underdeveloped areas with little or no current infrastructure, reliability sensitive applications in areas where wired infrastructure is unreliable, and political environments where monopolistic wired service providers maintain artificially high prices. Even in metropolitan areas and highly developed countries, wireless communication systems may be used for low-cost ubiquitous communication, new flexible data services, and emergency communication systems. In general, wireless communication systems may be used for voice communications just like conventional telephone systems, and for data communications in a radio-based wide area or local area network as well.

0003Wireless users access wireless communication systems using remote terminals such as cellular telephones and data modems equipped with radio transceivers. Such systems (and in particular the remote terminals) have protocols for initiating calls, receiving calls, and general transfer of information. The information transfer can be performed in real-time such as is the case for circuit-switched voice conversations and faxes, or in a store-and-forward manner such as is often the case for electronic mail, paging and other similar message transfer systems.

0004Wireless communication systems are generally allocated a portion of the radio frequency spectrum for their operation. The allocated portion of the spectrum is divided up into communication channels. These channels may be distinguished by frequency, by time, by code, or by some combination of the above. Each of these communication channels will be referred to herein as <i>conventional channels.</i> Depending on the available frequency allocations, the wireless system might have from one to several hundred communication channels. To provide full-duplex communication links, typically some of the communication channels are used for communication from base stations to users' remote terminals (the downlink), and others are used for communication from users' remote terminals to base stations (the uplink).

0005Wireless communication systems generally have one or more radio base stations, each of which provide coverage to a geographic area known as a cell and often serve as a point-of-presence (PoP) providing connection to a wide area network such as a Public Switched Telephone Network (PSTN). Often a predetermined subset of the available communication channels is assigned to each radio base station in an attempt to minimize the amount of interference experienced by users of the system. Within its cell, a radio base station can communicate simultaneously with many remote terminals by using different conventional communication channels for each remote terminal. As aforementioned, base stations can act as PoPs, providing connection to one or more wired communication systems. Such systems include local data networks, wide area data networks, and PSTNs. Thus, remote users are provided access to local and/or wide area data services and the local public telephone system. Base stations can also be used to provide local connectivity without direct access to a wired network such as in local area emergency and mobile battlefield communication systems. Base stations can provide connectivity of various kinds as well. In the aforementioned examples, point-to-point communications where roughly equal amounts of information flow in both directions between two users were assumed. In other applications such as interactive television, information is broadcast to all users simultaneously, and responses from many of the remote units arc to be processed at the base stations.

0006However, conventional wireless communication systems are comparatively spectrally inefficient. In conventional wireless communication systems, only one remote terminal can use any one conventional channel within a cell at any one time. If more than one remote terminal in a cell attempts to use the same channel at the same time, the downlink and uplink signals associated with the remote terminals interfere with each other. Since conventional receiver technology can not eliminate the interference in these combined uplink and downlink signals, remote terminals are unable to communicate effectively with the base station when interference is present. Thus, the total capacity of the system is limited by the number of conventional channels the base station has available, and in the overall system, by the way in which these channels are re-used among multiple cells. Consequently, conventional wireless systems arc unable to provide capacity anywhere near that of wired communication systems.

0007<patcit id="pcit0001" dnum="WO9312590A"><text>WO93/12590</text></patcit> discloses a base station with an array of antennas for receiving multiple signals transmitted in the same channel by a plurality of users. Parameters of the multiple signals are calculated and used to obtain the positions and velocities of the users. The locations and other related signal parameters are used to calculate appropriate spatial demultiplexing strategies, reconstructing the individual transmitted signals from the receiver measurements and reducing interference to acceptable levels. This hereto unavailable location information is used in solving the hand-off problem, and is also used to calculate an appropriate spatial multiplexing strategy for simultaneous transmission of signals to users in the same channel.

SUMMARY OF THE INVENTION

0008One aspect of the present invention provides a remote terminal for communicating with a base station in a wireless communication system, the remote terminal comprising: at least one antenna; receiving means for receiving transmissions from the base station on a conventional channel, wherein the conventional channel comprises any one of or any combination of a frequency channel, a time slot in a time division multiplexed system, and a code in a code division multiplexed system, the remote terminal being configured to receive calibration signals transmitted as predetermined waveforms from a spatial signature processor associated with the base station; processing means for computing a transmit spatial signature using the received calibration signals and the predetermined transmitted waveforms corresponding to the calibration signals, said transmit spatial signature characterising how the remote terminal receives signals from each of the antenna array elements at the base station in the conventional channel; and transmitting means for transmitting the computed transmit spatial signature to said base station.

0009A further aspect of the present invention provides a base station for transmitting to a plurality of remote terminals in a wireless communications system using a common conventional downlink channel, wherein said conventional channel comprises any one of or any combination of a frequency channel, a time slot in a time division multiplexed system, and a code in a code division multiplexed system, the base station comprising: transmission means including a plurality of transmit antenna elements and transmitters for transmitting multiplexed downlink signals to said plurality of remote terminals, and for transmitting predetermined calibration signals to a remote terminal; receiving means for receiving a transmit spatial signature determined by the remote terminal using the received calibration signals and transmitted to the base station by the remote terminal, said transmit spatial signature characterising how the remote terminal receives signals from each of the antenna array elements at the base station in the conventional channel; transmit spatial processing means for storing transmit spatial signatures for said plurality of remote terminals; and spatial multiplexing means using said transmit spatial signatures and downlink signals to produce said multiplexed downlink signals.

0010A further aspect of the invention provides a method at a remote terminal for communicating with a base station in a wireless communication system, the method comprising: receiving transmissions from the base station on a conventional channel, wherein the conventional channel comprises any one of or any combination of a frequency channel, a time slot in a time division multiplexed system, and a code in a code division multiplexed system, the remote terminal being configured to receive calibration signals transmitted as predetermined waveforms from a spatial signature processor associated with the base station; computing a transmit spatial signature using the received calibration signals and the predetermined transmitted waveforms corresponding to the calibration signals, said transmit spatial signature characterising how the remote terminal receives signals from each of the antenna array elements at the base station in the conventional channel; and transmitting the computed transmit spatial signature to said base station.

0011A futher aspect of the invention provides a method a method at a base station, for transmitting to a plurality of remote terminals in a wireless communications system using a common conventional downlink channel, wherein said conventional channel comprises any one of or any combination of a frequency channel, a time slot in a time division multiplexed system, and a code in a code division multiplexed system, the method comprising: transmitting multiplexed downlink signals to said plurality of remote terminals, and transmitting predetermined calibration signals to a remote terminal; receiving a transmit spatial signature determined by the remote terminal using the received calibration signals and transmitted to the base station by the remote terminal, said transmit spatial signature characterising how the remote terminal receives signals from each of the antenna array elements at the base station in the conventional channel; storing transmit spatial signatures for said plurality of remote terminals; and using said transmit spatial signatures and downlink signals to produce said multiplexed downlink signals.

0012Accordingly, an object of some embodiments of the present invention is to use antenna arrays and signal processing to separate combinations of received (uplink) signals. Another object of some embodiments of the present invention is to transmit spatially multiplexed downlink signals. The result is a dramatic increase in spectral efficiency, capacity, signal quality, and coverage of wireless communication systems. Capacity is increased by allowing multiple users to simultaneously share the same conventional communication channel within a cell without interfering with one another, and further by allowing more frequent reuse of the same conventional channel within geographic area covering many cells. Signal quality and coverage area are improved through intelligent processing of signals received from and transmitted by multiple antenna elements. Moreover, a further object of some embodiments of the present invention is to provide capacity gains by dynamically allocating conventional channels among base stations and remote terminals.

0013Briefly, embodiments of the invention comprise antenna arrays and signal processing means for measuring, calculating, storing, and using spatial signatures of receivers and transmitters in wireless communication systems to increase system capacity, signal quality, and coverage, and to reduce overall system cost The antenna array and signal processing means can be employed at base stations (PoPs) and remote terminals. Generally there can be different processing requirements at base stations where many signals are being concentrated than at remote terminals where in general only a limited number of communication links are being managed.

0014As an example, in a wireless local loop application, a particular base station might serve as a PoP many remote terminals and employ the antenna array and signal processing described herein. Additionally, remote terminals could employ antenna arrays and signal processing to further improve their capacity and signal quality over simpler remote terminals that handle fewer communication links. Herein, the distinction between base stations and remote terminals is that base stations generally act as concentrators connecting to multiple remote units simultaneously, possibly providing a high capacity connection to a wide area network. While for the sake of clarity much of the ensuing discussion is couched in terms of simple remote terminals that do not employ antenna arrays, nothing herein should be interpreted as preventing such an application. Thus, while hereafter spatial signatures will be associated primarily with remote terminals, when antenna arrays are employed at remote terminals, base stations will have associated spatial signatures as well.

0015Briefly, there are two spatial signatures associated with each remote terminal/base station pair on a particular frequency channel, where for the purpose of this discussion it is assumed that only base stations have antenna arrays. Base stations associate with each remote terminal in their cell a spatial signature related to how that remote terminal receives signals transmitted to it by the base station's antenna array, and a second spatial signature related to how the base station's receive antenna array receives signals transmitted by the remote terminal. In a system with many conventional channels, each remote terminal/base station pair has transmit and receive spatial signatures for each conventional channel.

0016The receive spatial signature characterizes how the base station antenna array receives signals from the particular remote unit in a particular conventional channel. In one embodiment, it is a complex vector containing responses (amplitude and phase with respect to a reference) of each the antenna element receivers, <i>i.e..</i> for an m-element array, <maths id="math0001" num="(1)"><math display="block"><msub><mi>a</mi><mi mathvariant="italic">br</mi></msub><mo>=</mo><msup><mfenced open="[" close="]"><msub><mi>a</mi><mi mathvariant="italic">br</mi></msub><mfenced><mn>1</mn></mfenced><mo>,</mo><msub><mi>a</mi><mi mathvariant="italic">br</mi></msub><mfenced><mn>2</mn></mfenced><mo>,</mo><mo>…</mo><mo>,</mo><msub><mi>a</mi><mi mathvariant="italic">br</mi></msub><mfenced><mi>m</mi></mfenced></mfenced><mi>T</mi></msup><mo>,</mo></math><img file="EP0804858B1_D0001.tif" /></maths> where a<i>_br</i>(<i>i</i>) is the response of the <i>i^th</i> receiver to a unit power transmitted signal from the remote terminal. Assuming that a narrowband signal <i>s_r</i>(<i>t</i>) is transmitted from the remote terminal, the base station receiver outputs at time <i>t</i> are then given by <maths id="math0002" num="(2)"><math display="block"><msub><mi>z</mi><mi>b</mi></msub><mfenced><mi>t</mi></mfenced><mo>=</mo><msub><mi>a</mi><mi mathvariant="italic">br</mi></msub><mo>⁢</mo><msub><mi>s</mi><mi>r</mi></msub><mo>⁢</mo><mfenced><mi>t</mi><mo>-</mo><mi>τ</mi></mfenced><mo>+</mo><msub><mi mathvariant="normal">n</mi><mi>b</mi></msub><mfenced><mi>t</mi></mfenced></math><img file="EP0804858B1_D0002.tif" /></maths> where r accounts for the mean propagation delay between the remote terminal and the base station antenna array, and n<i>_b</i>(<i>t</i>) represents noise present in the environment and the receivers.

0017The transmit spatial signature characterizes how the remote terminal receives signals from each of the antenna array elements at the base station in a particular conventional channel. In one embodiment, it is a complex vector containing relative amounts (amplitude and phase with respect to a reference) of each the antenna element transmitter outputs that are contained in the remote terminal receiver output. <i>i.e</i>., for an <i>m</i>-element array, <maths id="math0003" num="(3)"><math display="block"><msub><mi>a</mi><mi mathvariant="italic">rb</mi></msub><mo>=</mo><mfenced open="[" close="]"><msub><mi>a</mi><mi mathvariant="italic">rb</mi></msub><mfenced><mn>1</mn></mfenced><mo>,</mo><msub><mi>a</mi><mi mathvariant="italic">rb</mi></msub><mfenced><mn>2</mn></mfenced><mo>,</mo><mo>…</mo><mo>,</mo><msub><mi>a</mi><mi mathvariant="italic">rb</mi></msub><mfenced><mi>m</mi></mfenced></mfenced><mo>,</mo></math><img file="EP0804858B1_D0003.tif" /></maths> where a<i>_rb</i>(<i>i</i>) is the amplitude and phase (with respect to some fixed reference) of the remote terminal receiver output for a unit power signal transmitted from the <i>i^th</i> element in the base station array. Assuming that a vector of complex signals s<i>_b</i> = [s<i>_b</i>(1)....,s<i>_b</i>(<i>m</i>)]<i>^T</i> were transmitted from the antenna array, the output of the remote terminal receiver would be given by <maths id="math0004" num="(4)"><math display="block"><msub><mi>z</mi><mi>r</mi></msub><mfenced><mi>t</mi></mfenced><mo>=</mo><msub><mi>a</mi><mi mathvariant="italic">rb</mi></msub><mo>⁢</mo><msub><mi>s</mi><mi>b</mi></msub><mo>⁢</mo><mfenced><mi>t</mi><mo>-</mo><mi>τ</mi></mfenced><mo>+</mo><msub><mi>n</mi><mi>r</mi></msub><mfenced><mi>t</mi></mfenced><mo>,</mo></math><img file="EP0804858B1_D0004.tif" /></maths> where <i>n_r</i>(<i>t</i>) represents noise present in the environment and the receiver. These spatial signatures are calculated (estimated) and stored at each base station for each remote terminal in its cell and for each conventional channel. For fixed remote terminals and base stations in stationary environments, the spatial signatures can be updated infrequently. In general, however, changes in the RF propagation environment between the base station and the remote terminal can alter the signatures and require that they be updated. Note that henceforth, the time argument in parentheses will be suppressed: integers inside parentheses will be used solely for indexing into vectors and matrices.

0018In the previous discussion, temporally matched receivers and transmitters were assumed. If there are differences in the temporal responses, these can be equalized using temporal filtering techniques as is well-known. Furthermore, the channel bandwidths were assumed to be small compared to the center frequency of operation. Large bandwidth channels may require more than one complex vector to accurately describe the outputs as is well known.

0019When more than one remote terminal wants to communicate at the same time, the signal processing means at the base station uses the spatial signatures of the remote terminals to determine if subsets of them can communicate with the base station simultaneously by sharing a conventional channel. In a system with m receive and m transmit antenna elements, up to <i>m</i> remote terminals can share the same conventional channel at the same time.

0020When multiple remote terminals are sharing a single conventional uplink channel, the multiple antenna elements at the base station each measure a combination of the arriving uplink signals and noise. These combinations result from the relative locations of the antenna elements, the locations of the remote terminals, and the RF propagation environment. The signal processing means calculates spatial demultiplexing weights to allow the uplink signals to be separated from the combinations of uplink signals measured by the multiple antenna elements.

0021In applications where different downlink signals are to be sent from the base station to the remote terminals, the signal processing means computes spatial multiplexing weights that are used to produce multiplexed downlink signals, which when transmitted from the antenna elements at the base station result in the correct downlink signal being received at each remote terminal with appropriate signal quality.

0022In applications where the same signal is to be transmitted from the base station to a large number (mere than the number of antenna elements) of remote terminals the signal processing means computes weights appropriate for broadcasting the signal, covering the area necessary to reach all the remote terminals.

0023Therefore the signal processing means facilitates simultaneous communication between a base station and multiple remote terminals on the same conventional channel. The conventional channel may be a frequency channel, a time slot in a time division multiplexed system, a code in a code division multiplexed system, or any combination of the above.

0024In one embodiment, all elements of a single antenna array transmit and receive radio frequency signals, while in another embodiment the antenna array includes separate transmit antenna elements and receive antenna elements. The number of transmit and receive elements need not be the same. If they are not the same, the maximum number of point-to-point links that can simultaneously be established in one conventional channel is given by the smaller of the two numbers. Embodiments of the invention and objects and features thereof will be more readily apparent from the following detailed description together with the figures and appended claims.

Brief Description of The Drawings

0025<ul id="ul0001" list-style="none" compact="compact"><li><figref idref="f0001">Figure 1</figref> is a functional block diagram of a base station in accordance with the invention.</li><li><figref idref="f0002">Figure 2</figref> is a functional block diagram of multichannel receivers in the base station.</li><li><figref idref="f0003">Figure 3</figref> is a functional block diagram of a spatial demultiplexer in the base station.</li><li><figref idref="f0005">Figure 5</figref> is a functional block diagram of a multichannel transmitter in the base station.</li><li><figref idref="f0006">Figure 6</figref> is a functional block diagram of a spatial processor in the base station.</li><li><figref idref="f0007">Figure 7</figref> is a functional block diagram of a remote terminal with a transponder switch.</li><li><figref idref="f0008">Figure 8</figref> is a functional block diagram of a remote terminal.</li><li><figref idref="f0009">Figure 9</figref> is a schematic diagram of a network system comprised of three base stations and a multiple base station controller.</li></ul>

List of Reference Numerals

0026<dl id="dl0001" compact="compact"><dt>1.</dt><dd>base station</dd><dt>2.</dt><dd>base station communication link</dd><dt>3.</dt><dd>base station controller</dd><dt>4.</dt><dd>demodulated received signal</dd><dt>5.</dt><dd>spatially separated uplink signals</dd><dt>6.</dt><dd>received signal measurements</dd><dt>7.</dt><dd>demultiplexing weights</dd><dt>8.</dt><dd>data to be transmitted directionally</dd><dt>9.</dt><dd>modulated signal to be multiplexed for transmission</dd><dt>10.</dt><dd>modulated, spatially multiplexed signals to be transmitted</dd><dt>11.</dt><dd>calibration signals to be transmitted</dd><dt>12.</dt><dd>multiplexing weights</dd><dt>13.</dt><dd>spatial processor</dd><dt>14.</dt><dd>multichannel transmitters</dd><dt>15.</dt><dd>multichannel receivers</dd><dt>16a.</dt><dd>multichannel receiver</dd><dt>16m.</dt><dd>multichannel receiver</dd><dt>17a.</dt><dd>multichannel transmitter</dd><dt>17m.</dt><dd>multichannel transmitter</dd><dt>18a.</dt><dd>transmit antenna</dd><dt>18m.</dt><dd>transmit antenna</dd><dt>19a.</dt><dd>receive antenna</dd><dt>19m.</dt><dd>receive antenna</dd><dt>20.</dt><dd>spatial demultiplexer</dd><dt>21.</dt><dd>adder</dd><dt>22a.</dt><dd>multipliers</dd><dt>22m.</dt><dd>multipliers</dd><dt>23.</dt><dd>spatial multiplexer</dd><dt>24.</dt><dd>signal modulator</dd><dt>25.</dt><dd>signal demodulator</dd><dt>26a.</dt><dd>multipliers</dd><dt>26m.</dt><dd>multipliers</dd><dt>27.</dt><dd>spatial control data</dd><dt>28.</dt><dd>spatial parameter data</dd><dt>29.</dt><dd>common receiver oscillator</dd><dt>30.</dt><dd>receiver control data</dd><dt>31.</dt><dd>transmitter control data</dd><dt>32.</dt><dd>common transmitter oscillator</dd><dt>33.</dt><dd>spatial processor controller</dd><dt>34.</dt><dd>active remote terminal list</dd><dt>35.</dt><dd>channel selector</dd><dt>36.</dt><dd>remote terminal database</dd><dt>37.</dt><dd>spatial weight processor</dd><dt>38.</dt><dd>spatial signature processor</dd><dt>39.</dt><dd>remote terminal antenna</dd><dt>40.</dt><dd>remote terminal duplexer</dd><dt>41.</dt><dd>remote terminal duplexer output</dd><dt>42.</dt><dd>remote terminal receiver</dd><dt>43.</dt><dd>remote terminal received signal</dd><dt>44.</dt><dd>remote terminal received calibration signal</dd><dt>45.</dt><dd>remote terminal demodulator</dd><dt>46.</dt><dd>remote terminal demodulated data</dd><dt>47.</dt><dd>remote terminal keyboard and keyboard controller</dd><dt>48.</dt><dd>remote terminal keyboard data</dd><dt>49.</dt><dd>remote terminal display data</dd><dt>50.</dt><dd>remote terminal display and display controller</dd><dt>51.</dt><dd>remote terminal modulator</dd><dt>52.</dt><dd>remote terminal data to be transmitted</dd><dt>53.</dt><dd>remote terminal modulated data to be transmitted</dd><dt>54.</dt><dd>remote terminal transmitter</dd><dt>55.</dt><dd>remote terminal transmitter output</dd><dt>56.</dt><dd>remote terminal transmitter control data</dd><dt>57.</dt><dd>remote terminal receiver control data</dd><dt>58.</dt><dd>remote terminal microphone</dd><dt>59.</dt><dd>remote terminal microphone signal</dd><dt>60.</dt><dd>remote terminal speaker</dd><dt>61.</dt><dd>remote terminal speaker signal</dd><dt>62.</dt><dd>remote terminal central processing unit</dd><dt>63.</dt><dd>remote terminal transponder switch</dd><dt>64.</dt><dd>remote terminal transponder switch control</dd><dt>65.</dt><dd>wide area network</dd><dt>66.</dt><dd>multiple base station controller</dd><dt>67a.</dt><dd>cell boundary</dd><dt>67b.</dt><dd>cell boundary</dd><dt>67c.</dt><dd>cell boundary</dd><dt>68.</dt><dd>high speed message link</dd><dt>69.</dt><dd>remote terminal</dd></dl>

Description of Invention

0027<figref idref="f0001">Figure 1</figref> depicts the preferred embodiment of a base station 1. A base station controller 3 acts as an interface between base station 1 and any external connection via a base station communication link 2, and serves to coordinate the overall operation of base station 1. In the preferred embodiment, base station controller 3 is implemented with a conventional central processing unit and associated memory and programming.

0028Incoming or uplink radio transmissions impinge on an antenna array composed of a number, <i>m</i>, of receive antenna elements 19(a,... ,m) each of whose outputs is connected to one of m multichannel receivers in a bank of phase-coherent multichannel receivers 15. Multichannel receivers 15 have well-matched amplitude and phase responses across the frequency bands of interest, or, as is well known, correction filters are implemented to account for any differences.

0029The illustrative embodiment describes a conventional frequency division multiple access system. Each multichannel receiver is capable of handling multiple frequency channels. The symbol <i>N_cc</i> will be used to reference the maximum number of conventional frequency channels that can be handled by the receivers. Depending on the frequencies allocated for the operation of the wireless communication system and the bandwidths chosen for particular communication links. <i>N_cc</i> could be as small as one (a single frequency channel) or as large as thousands. In alternate embodiments, multichannel receivers 15 might instead handle multiple time slots, multiple codes, or some combination of these well known multiple access techniques.

0030In each conventional channel, receive antenna elements 19(a,....m) each measure a combination of the arriving uplink signals from the remote terminals sharing this conventional channel. These combinations result from the relative locations of the antenna elements, the locations of the remote terminals, and the RF propagation environment, and for narrowband signals are given by equation (2).

0031<figref idref="f0002">Figure 2</figref> depicts individual multichannel receivers 16(a, ...,m) with antenna element connections, common local receiver oscillators 29, one for each conventional frequency channel to be used at that base station, and received signal measurements 6. Common local receiver oscillators 29 ensure that the signals from receive antenna elements 19(a, ....m) are coherently down-converted to baseband: its <i>N_cc</i> frequencies are set so that multichannel receivers 16(a, ....m) extract all <i>N_cc</i> frequency channels of interest. The frequencies of common local receiver oscillators 29 are controlled by a spatial processor 13 (<figref idref="f0001">figure 1</figref>) via receiver control data 30. In an alternate embodiment, where multiple frequency channels are all contained in a contiguous frequency band, a common local oscillator is used to downconvert the entire band which is then digitized, and digital filters and decimators extract the desired subset of channels using well known techniques.

0032The illustrative embodiment describes a frequency division multiple access system. In a time division multiple access or code division multiple access system, common oscillators 29 would be augmented to relay common time slot or common code signals respectively from spatial processor 13. via receiver control data 30. to multichannel receivers 16(a.... ,m). In these embodiments, multichannel receivers 16(a....,m) perform selection of conventional time division channels or conventional code division channels in addition to down conversion to baseband.

0033Referring again to <figref idref="f0001">Figure 1</figref>. multichannel receivers 15 produce received signal measurements 6 which are supplied to spatial processor 13 and to a set of spatial demultiplexers 20. In this embodiment, received signal measurements 6 contain m complex baseband signals for each of <i>N_cc</i> frequency channels.

0034<figref idref="f0006">Figure 6</figref> shows a more detailed block diagram of spatial processor 13. Spatial processor 13 produces and maintains spatial signatures for each remote terminal for each conventional frequency channel, and calculates spatial multiplexing and demultiplexing weights for use by spatial demultiplexers 20 and spatial multiplexers 23. In the preferred embodiment, spatial processor 13 is implemented using a conventional central processing unit. Received signal measurements 6 go into a spatial signature processor 38 which estimates and updates spatial signatures. Spatial signatures are stored in a spatial signature list in a remote terminal database 36 and are used by channel selector 35 and spatial weight processor 37, which also produces demultiplexing weights 7 and multiplexing weights 12. A spatial processor controller 33 connects to spatial weight processor 37 and also produces receiver control data 30 transmitter control data 31 and spatial control data 27.

0035Referring again to <figref idref="f0001">Figure 1</figref>. spatial demultiplexers 20 combine received signal measurements 6 according to spatial demultiplexing weights 7. <figref idref="f0003">Figure 3</figref> shows a spatial demultiplexer 20 for a single conventional channel. In this embodiment, arithmetic operations in spatial demultiplexer 20 are carried out using general purpose arithmetic chips. In <figref idref="f0003">figure 3</figref>, <i>z_b</i>(<i>i</i>) denotes the <i>i^th</i> component of received signal measurement vector 6 for a single conventional channel, and <maths id="math0005"><math display="inline"><msubsup><mi mathvariant="normal">w</mi><mi mathvariant="italic">rx</mi><mo>*</mo></msubsup><mfenced><mi>i</mi></mfenced></math><img file="EP0804858B1_D0005.tif" /></maths> denotes the complex conjugate of the i^th component of the spatial demultiplexing weight vector 7 for a remote terminal using this conventional channel.

0036For each remote terminal on each conventional channel, the spatial demultiplexer 20 computes the inner-product of the spatial demultiplexing weights for the conventional channel with the received signal measurements 6: <maths id="math0006" num="(5)"><math display="block"><msubsup><mi mathvariant="normal">w</mi><mi mathvariant="italic">rx</mi><mo>*</mo></msubsup><mo>⁢</mo><msub><mi mathvariant="normal">z</mi><mi>b</mi></msub><mo>=</mo><msubsup><mi mathvariant="normal">w</mi><mi mathvariant="italic">rx</mi><mo>*</mo></msubsup><mfenced><mi>i</mi></mfenced><mo>⁢</mo><msub><mi mathvariant="normal">z</mi><mi>b</mi></msub><mfenced><mn>1</mn></mfenced><mo>+</mo><mo>…</mo><mo>+</mo><msubsup><mi mathvariant="normal">w</mi><mi mathvariant="italic">rx</mi><mo>*</mo></msubsup><mfenced><mi>m</mi></mfenced><mo>⁢</mo><msub><mi mathvariant="normal">z</mi><mi>b</mi></msub><mfenced><mi>m</mi></mfenced><mo>,</mo></math><img file="EP0804858B1_D0006.tif" /></maths> where (·)* indicates complex conjugation, numbers inside parentheses indicate element number (<i>e.g.,</i> w<i>_rx</i>(<i>i</i>) is the <i>i^th</i> component of the vector w<i>_rx</i>), the multiplication is performed by multipliers 22(a, ....m), and the addition is performed by adder 21. For each remote terminal on each conventional channel, the output of adder 21 given by equation (5) comprises the spatially separated uplink signals 5.

0037Referring again to <figref idref="f0001">Figure 1</figref>. the outputs of spatial demultiplexers 20 are spatially separated uplink signals 5 for each remote terminal communicating with the base station. Spatially separated uplink signals 5 are demodulated by signal demodulators 25, producing demodulated received signals 4 for each remote terminal communicating with the base station. Demodulated received signals 4 and corresponding spatial control data 27 are available to base station controller 3.

0038In embodiments where channel coding of the signals sent by remote terminals is performed, base station controller 3 sends the demodulated received signals 4 to spatial processor 13 which, using well known decoding techniques, estimates Bit-Error-Rates (BERs) and compares them against acceptable thresholds stored in the remote terminal database 36. If the BERs are unacceptable, spatial processor 13 reallocates resources so as to alleviate the problem. In one embodiment, links with unacceptable BERs are assigned to new channels using the same strategy as adding a new user with the exception that the current channel is not acceptable unless the current set of users of that particular channel changes. Additionally, recalibration of the receive signature for that remote terminal/base station pair is performed when that conventional channel is available.

0039For transmission, signal modulators 24 produce modulated signals 9 for each remote terminal the base station is transmitting to, and a set of spatial multiplexing weights 12 for each remote terminal are applied to the respective modulated signals in spatial multiplexers 23 to produce spatially multiplexed signals to be transmitted 10 for each of the m transmit antennas 18(a.... ,m) and each of the <i>N_cc</i> conventional channels.

0040In the illustrative embodiment the number <i>N_cc</i> of downlink conventional channels is the same as the number <i>N_cc</i> of uplink conventional channels. In other embodiments, there may be different numbers of uplink and downlink conventional channels. Furthermore, the channels may be of different types and bandwidths as is the case for an interactive television application where the downlink is comprised of wideband video channels and the uplink employs narrowband audio/data channels.

0041Additionally, the illustrative embodiment shows the same number of antenna elements, <i>m</i>, for transmit and receive. In other embodiments, the number of transmit antenna elements and the number of receive antenna elements may be different, up to and including the case where transmit employs only one transmit antenna element in an omnidirectional sense such as in an interactive television application.

0042<figref idref="f0004">Figure 4</figref> shows the spatial multiplexer for one remote terminal on a particular conventional channel. Arithmetic operations in spatial multiplexer 23 are carried out using general purpose arithmetic chips. The component of modulated signals 9 destined for this remote terminal on this conventional channel is denoted by <i>s_b</i>, and W<i>_tx</i>(<i>i</i>) denotes the <i>i^th</i> component of spatial multiplexing weight vector 12 for this remote terminal on this conventional channel.

0043For each remote terminal on each conventional channel, the spatial multiplexer 23 computes the product of its spatial multiplexing weight vector (from the spatial multiplexing weights 12) with its modulated signal s<i>_b</i> (from the modulated signals 9): <maths id="math0007" num="(6)"><math display="block"><msubsup><mi mathvariant="normal">w</mi><mi mathvariant="italic">tx</mi><mo>*</mo></msubsup><mo>⁢</mo><msub><mi>s</mi><mi>b</mi></msub><mo>=</mo><mfenced open="[" close="]"><mtable><mtr><mtd><msubsup><mi mathvariant="normal">w</mi><mi mathvariant="italic">tx</mi><mo>*</mo></msubsup><mfenced><mn>1</mn></mfenced><mo>⁢</mo><msub><mi>s</mi><mi>b</mi></msub></mtd></mtr><mtr><mtd><mo>⋮</mo></mtd></mtr><mtr><mtd><msubsup><mi mathvariant="normal">w</mi><mi mathvariant="italic">tx</mi><mo>*</mo></msubsup><mfenced><mi>m</mi></mfenced><mo>⁢</mo><msub><mi>s</mi><mi>b</mi></msub></mtd></mtr></mtable></mfenced><mo>,</mo></math><img file="EP0804858B1_D0007.tif" /></maths> where (·)* indicates complex conjugate (transpose) and the multiplication is performed by multipliers 26(a, ....m). For each conventional channel, equation (6) is evaluated by the spatial multiplexer 23 for each remote terminal that is being transmitted to on this conventional channel. Corresponding to each remote terminal is a different multiplexing weight vector and modulated signal. For each conventional channel, spatial multiplexer 23 adds the spatially multiplexed signals for each remote terminal being transmitted to on this conventional channel, producing modulated and spatially multiplexed signals 10 that are the signals to be transmitted for each conventional downlink channel from each antenna.

0044Modulated and spatially multiplexed signals 10 are inputs to a bank of m phase coherent multichannel transmitters 14. Multichannel transmitters 14 either have well-matched amplitude and phase responses across the frequency bands of interest, or, as is well known, correction filters are implemented to account for any differences. <figref idref="f0005">Figure 5</figref> depicts multichannel transmitters 17(a, ...,m) with antenna connections, common local transmitter oscillators 32, and digital inputs 10. Common local transmitter oscillators 32 ensure that the relative phases of spatially multiplexed signals 10 are preserved during transmission by transmit antennas 18(a, ....m). The frequencies of common local transmitter oscillators 32 are controlled by spatial processor 13 (see <figref idref="f0001">figure 1</figref>) via transmitter control data 31.

0045In an alternate embodiment, spatial multiplexer 23 uses well known baseband multiplexing techniques to multiplex all the calculated conventional channel signals to be transmitted into a single wideband signal to be upconverted and transmitted by each of the multichannel transmitters 17(a....,m). The multiplexing can be performed either digitally or in analog as appropriate.

0046The illustrative embodiment shows a system with multiple frequency channels. In a time division multiple access or code division multiple access system, common oscillators 32 would be augmented to relay common time slot or common code signals respectively from spatial processor 13. via transmitter control data 31. to multichannel transmitters 17(a, ....m).

0047Referring again to <figref idref="f0001">figure 1</figref>. in applications where transmit spatial signatures are required, spatial processor 13 is also able to transmit predetermined calibration signals 11 for each antenna on a particular conventional downlink channel. Spatial processor 13 instructs multichannel transmitters 17(a....,m), via transmitter control data 31. to transmit predetermined calibration signals 11 in place of spatially multiplexed signals 10 for a particular conventional downlink channel. This is one mechanism used for determining the transmit spatial signatures of the remote terminals on this conventional downlink channel.

0048In alternate embodiments where well known channel coding techniques are used to encode the signals to be transmitted to remote terminals, remote terminals employ well known decoding techniques to estimate BERs which are then reported back to the base station on their uplink channel. If these BERs exceed acceptable limits, corrective action is taken. In one embodiment, the corrective active involves reallocating resources by using the same strategy as adding a new user with the exception that the current channel is not acceptable unless the current set of users of that particular channel changes. Additionally, recalibration of the transmit signature for that remote terminal/base station pair is performed when that conventional channel is available.

0049<figref idref="f0007">Figure 7</figref> depicts the component arrangement in a remote terminal that provides voice communication. The remote terminal's antenna 39 is connected to a duplexer 40 to permit antenna 39 to be used for both transmission and reception. In an alternate embodiment, separate receive and transmit antennas are used eliminating the need for duplexer 40. In another alternate embodiment where reception and transmission occur on the same frequency channel but at different times, a transmit/receive (TR) switch is used instead of a duplexer as is well known. Duplexer output 41 serves as input to a receiver 42. Receiver 42 produces a down-converted signal 43 which is the input to a demodulator 45. A demodulated received voice signal 61 is input to a speaker 60.

0050Demodulated received control data 46 is supplied to a remote terminal central processing unit 62 (CPU). Demodulated received control data 46 is used for receiving data from base station 1 during call setup and termination, and in an alternate embodiment, for determining the quality (BER) of the signals being received by the remote terminal for transmission back to the base station as described above.

0051Remote terminal CPU 62 is implemented with a standard microprocessor. Remote terminal CPU 62 also produces receiver control data 57 for selecting the remote terminal's reception channel, transmitter control data 56 for setting the remote terminal's transmission channel and power level, control data to be transmitted 52, and display data 49 for remote terminal display 50. Remote terminal CPU 62 also receives keyboard data 48 from remote terminal keyboard 47.

0052The remote terminal's voice signal to be transmitted 59 from microphone 58 is input to a modulator 51. Control data to be transmitted 52 is supplied by remote terminal CPU 62. Control data to be transmitted 52 is used for transmitting data to base station 1 during call setup and termination as well as for transmitting information during the call such as measures of call quality (<i>e.g.,</i> bit error rates (BERs)). The modulated signal to be transmitted 53. output by modulator 31. is up-converted and amplified by a transmitter 34. producing a transmitter output signal 55. Transmitter output 55 is then input to duplexer 40 for transmission by antenna 39.

0053In an alternate embodiment, the remote terminal provides digital data communication. Demodulated received voice signal 61, speaker 60, microphone 58, and voice signal to be transmitted 59 are replaced by digital interfaces well-known in the art that allow data to be transmitted to and from an external data processing device (for example, a computer).

0054Referring again to <figref idref="f0007">figure 7</figref>, the remote terminal allows received data 43 to be transmitted back to base station 1 via switch 63 controlled by remote terminal CPU 62 through switch control signal 64. In normal operation, switch 63 drives transmitter 54 with modulated signal 53 of modulator 51. When the remote terminal is instructed by base station 1 to enter calibration mode, remote terminal CPU 62 toggles switch control signal 64. which instructs switch 63 to drive transmitter 54 with received data 43.

0055<figref idref="f0008">Figure 8</figref> shows an alternate embodiment of the remote terminal calibration function. Switch 63 of <figref idref="f0007">figure 7</figref> is no longer used. Instead, the output of receiver 42 is supplied to remote terminal CPU 62 by data connection 44. In normal operation remote terminal CPU 62 ignores data connection 44. In calibration mode, remote terminal CPU 62 uses data connection 44 to compute the remote terminal's transmit spatial signature, which is transmitted back to base station 1 through modulator 51 and transmitter 54 as control data to be transmitted 52.

0056In an alternate embodiment, special calibration procedures in the remote terminal are not required. In many conventional wireless protocol standards, remote terminals regularly report received signal strength or receive signal quality back the base station. In this embodiment, the received signal strength reports are sufficient to compute the remote terminal's transmit spatial signature, as described below.

Operation of Invention

General Principles - Base Station

0057In many respects, the spectrally efficient base station shown in <figref idref="f0001">figure 1</figref> behaves much like a standard wireless communication system base station. The primary distinction is that the spectrally efficient base station supports many more simultaneous conversations than it has conventional communication channels. The conventional communication channels may be frequency channels, time channels, code channels, or any combination of these. The spatial multiplexer/demultiplexer increases the system capacity by allowing multiple spatial channels on each of these conventional channels. Moreover, by combining signals from multiple receive antennas, the spatial demultiplexer 20 produces spatially separated uplink signals 5 that have substantially improved signal-to-noise, reduced interference, and improved quality in multipath environments compared to a standard base station.

0058In the illustrative embodiment, a wireless communication system comprised of multiple remote terminals and base stations incorporating antenna arrays and spatial signal processing is described. Such systems have application, for example, in providing wireless access to the local PSTN. Information transfers (or calls) are initiated by either a remote terminal or by communication link 2 through base station controller 3. Call initialization takes place on a downlink and uplink control channel as is well known in the art. In the present embodiment, the downlink control channel is transmitted using transmission antennas 18(a,....m). In an alternate embodiment, the downlink control channel is broadcast from a single, omnidirectional antenna. Base station controller 3 passes the identification of the remote terminal to be involved in the call to spatial processor 13 which uses the stored spatial signatures of that remote terminal to determine which conventional communication channel the remote terminal should use. The selected channel may already be occupied by several remote terminals, however spatial processor 13 uses the spatial signatures of all of the remote terminals on that channel to determine that they can share the channel without interference. In a system with <i>m</i> receive and m transmit antenna elements, up to m remote terminals can share the same conventional channel. More generally, the number of point-to-point full-duplex communication links that can occupy the same conventional channel at the same time is given by the smaller of the number of receive and transmit elements.

0059Spatial processor 13 uses calculated spatial multiplexing and demultiplexing weights for the selected channel and the remote terminal in question to configure spatial multiplexer 23 and spatial demultiplexer 20. Spatial processor 13 then informs controller 3 of the selected channel. As in a conventional base station, controller 3 then commands the remote terminal (via the downlink control channel) to switch to the selected channel for continued communications. In the event that the remote terminal has power control capabilities, as is well known in the art, controller 3 also commands the remote terminal to adjust its power to an appropriate level based on parameters such as the power levels of the other remote terminals sharing the same conventional channel and the required signal quality for each link as discussed below. At the termination of communications, the remote terminal returns to its idle state where it monitors the downlink control channel awaiting its next call. This frees up that "spatial channel for another remote terminal.

Spatial Processing - Base Station

0060<figref idref="f0006">Figure 6</figref> shows a block diagram of spatial processor 13. It is controlled by spatial processor controller 33, which interfaces to base station controller 3 via link 27. Spatial processor controller 33 controls the gain and frequency settings of multichannel transmitters 14 and multichannel receivers 15 by control lines 31 and 30.

0061Spatial processor 13 maintains an active remote terminal list 34 that catalogs which remote terminals are currently using each conventional communication channel as well as their current transmit power levels. Other parameters of the remote terminals such as modulation formats currently used, receiver noise levels in current frequency channels, and current signal quality requirements are stored as well. Spatial processor 13 also maintains a spatial signature list in the remote terminal database 36. which in alternate embodiments includes remote terminals' power control levels, allowed conventional frequency channels for receive and transmit, and list of modulation formats understood.

0062The spatial signature list in the remote terminal database 36 contains a transmit spatial signature, a<i>_rb,</i> and a receive spatial signature, a<i>_br</i>, for every frequency of operation for each remote terminal. In another embodiment, estimates of the quality (<i>e.g</i>., estimate error covariances) of the spatial signatures are stored as well. As aforementioned, the transmit spatial signature, a<i>_rb,</i> for a particular remote terminal and a particular frequency channel is defined as the vector of relative complex signal amplitudes that would be seen arriving at that particular remote terminal as a result of identical (equal amplitude and phase) unit power narrow band signals, at that particular frequency, being transmitted through multichannel transmitters 14 and transmission antennas 18(a, ....m). The transmit spatial signature includes the effects of the propagation environment between the base station and the remote terminal, as well as any amplitude and phase differences in multichannel transmitters 14. antenna cables, and transmission antennas 18(a. ... ,m). The receive spatial signature, a<i>_br</i>, for a particular remote terminal and a particular frequency channel, is defined as the vector of complex signal amplitudes that would be measured at the outputs of multichannel receiver 16 given a single unit power narrow band signal being transmitted by that particular remote terminal, at that particular frequency.

0063When the base station controller 1 forwards a call initialization request for a particular remote terminal via link 27. a channel selector 35 searches active remote terminal list 34 to find a conventional communication channel that can accommodate the remote terminal. In the preferred embodiment, there is a receive active remote terminal list and a transmit active remote terminal list which are used by channel selector 35 in forming both a multiplexing and a demultiplexing spatial signature matrix for each conventional channel. For each conventional channel, the columns of the demultiplexing and rows of the multiplexing spatial signature matrices are the stored receive and transmit spatial signatures of each of the remote terminals currently active on (using) that channel plus one more column containing the appropriate spatial signature of the remote terminal requesting a communication channel.

0064The multiplexing spatial signature matrix for each channel, <i>A_rb.p</i> (where p denotes the conventional channel number), is formed using transmit spatial signatures as shown in equation (7): <maths id="math0008" num="(7)"><math display="block"><msub><mi>A</mi><mrow><mi mathvariant="italic">rb</mi><mo mathvariant="italic">,</mo><mi mathvariant="italic">p</mi></mrow></msub><mo>=</mo><mfenced open="[" close="]"><mtable><mtr><mtd><msubsup><mi mathvariant="normal">a</mi><mrow><mi mathvariant="italic">rb</mi><mo>,</mo><mi>p</mi></mrow><mn>1</mn></msubsup></mtd></mtr><mtr><mtd><mo>⋮</mo></mtd></mtr><mtr><mtd><msubsup><mi mathvariant="normal">a</mi><mrow><mi mathvariant="italic">rb</mi><mo mathvariant="italic">,</mo><mi mathvariant="italic">p</mi></mrow><msub><mi>n</mi><mi>p</mi></msub></msubsup></mtd></mtr></mtable></mfenced><mo>,</mo></math><img file="EP0804858B1_D0008.tif" /></maths> where <maths id="math0009"><math display="inline"><mtable><mtr><mtd><msubsup><mi mathvariant="normal">a</mi><mrow><mi mathvariant="italic">rb</mi><mo>,</mo><mi>p</mi></mrow><mi>i</mi></msubsup></mtd></mtr></mtable></math><img file="EP0804858B1_D0009.tif" /></maths> is the transmit spatial signature for <i>i</i>^th remote terminal assigned to channel <i>p</i> and <i>n_p</i> is the total number of remotes on conventional channel p.

0065The demultiplexing spatial signature matrix, <i>A_br,p,</i> is formed using receive spatial signatures as shown in equation (8): <maths id="math0010" num="(8)"><math display="block"><msub><mi>A</mi><mrow><mi mathvariant="italic">br</mi><mo>,</mo><mi>p</mi></mrow></msub><mfenced open="[" close="]"><mtable><mtr><mtd><msubsup><mi mathvariant="normal">a</mi><mrow><mi mathvariant="italic">br</mi><mo>,</mo><mi>p</mi></mrow><mn>1</mn></msubsup></mtd></mtr></mtable><mtable><mtr><mtd><msubsup><mi mathvariant="normal">a</mi><mrow><mi mathvariant="italic">br</mi><mo>,</mo><mi>p</mi></mrow><mn>2</mn></msubsup></mtd></mtr></mtable><mo>⋯</mo><mtable><mtr><mtd><msubsup><mi mathvariant="normal">a</mi><mrow><mi mathvariant="italic">br</mi><mo>,</mo><mi>p</mi></mrow><msub><mi>n</mi><mi>p</mi></msub></msubsup></mtd></mtr></mtable></mfenced><mo>,</mo></math><img file="EP0804858B1_D0010.tif" /></maths> where <maths id="math0011"><math display="inline"><mtable><mtr><mtd><msubsup><mi mathvariant="normal">a</mi><mrow><mi mathvariant="italic">br</mi><mo>,</mo><mi>p</mi></mrow><mi>i</mi></msubsup></mtd></mtr></mtable></math><img file="EP0804858B1_D0011.tif" /></maths> is the receive spatial signature for <i>i^th</i> remote terminal assigned to channel <i>p</i>.

0066Channel selector 35 calculates functions of these signature matrices to assess whether or not communication between the base station and the new remote terminal can be successfully carried out on the selected conventional channel. In the preferred embodiment, channel selector 35 first calculates spatial multiplexing and demultiplexing weights for that remote terminal and then uses these weights to estimate link performance.

0067In the illustrative embodiment, spatial multiplexing weights are the rows of a matrix <i>W_tx</i> given in equation (9): <maths id="math0012" num="(9)"><math display="block"><msub><mi>W</mi><mi mathvariant="italic">tx</mi></msub><mo>=</mo><msup><mrow><msub><mi>S</mi><mi>b</mi></msub><mfenced><msub><mi>A</mi><mi mathvariant="italic">rb</mi></msub><mo>⁢</mo><msubsup><mi>A</mi><mi mathvariant="italic">rb</mi><mo>*</mo></msubsup></mfenced></mrow><mrow><mo>-</mo><mn>1</mn></mrow></msup><mo>⁢</mo><msub><mi>A</mi><mi mathvariant="italic">rb</mi></msub><mo>,</mo></math><img file="EP0804858B1_D0012.tif" /></maths> where (·)^-1 is the inverse of a matrix, (·)* is the complex conjugate transpose of a matrix. <i>A_rb</i> is the multiplexing spatial signature matrix <i>A_rb,p</i> associated with the relevant conventional channel, and <i>S_b</i> is a (diagonal) matrix containing the amplitudes of the signals to be transmitted. The amplitudes to be transmitted. <i>S_b,</i> are calculated in the preferred embodiment using the (diagonal) matrix of remote terminal receiver mean-square noise voltages (<i>N</i>) and the diagonal matrix of minimum desired signal qualities (<i>SNR_des</i>) as given in equation (10): <maths id="math0013" num="(10)"><math display="block"><msub><mi>S</mi><mi>b</mi></msub><mo>=</mo><msup><mfenced><msub><mi mathvariant="italic">SNR</mi><mi mathvariant="italic">des</mi></msub><mo>×</mo><mi>N</mi></mfenced><mrow><mn>1</mn><mo>/</mo><mn>2</mn></mrow></msup></math><img file="EP0804858B1_D0013.tif" /></maths> Now channel selector 35 calculates the average mean-square voltage (power) <i><o ostyle="single">P_b</o></i> to be transmitted from each element as the sum of squares of the elements m each row of <i>W_tx. i.e..</i><maths id="math0014" num="(11)"><math display="block"><mover><msub><mi>P</mi><mi>b</mi></msub><mo>‾</mo></mover><mo>=</mo><mi>diag</mi><mfenced open="{" close="}"><msub><mi>W</mi><mi mathvariant="italic">tx</mi></msub><mo>⁢</mo><msubsup><mi>W</mi><mi mathvariant="italic">tx</mi><mo>*</mo></msubsup></mfenced><mn>.</mn></math><img file="EP0804858B1_D0014.tif" /></maths> and the peak square voltage (power) <maths id="math0015"><math display="inline"><msubsup><mi>P</mi><mi>b</mi><mi mathvariant="italic">peak</mi></msubsup></math><img file="EP0804858B1_D0015.tif" /></maths> to be transmitted from each element as the square of sum of the magnitude of the elements in each row of <i>W_tx, i.e</i>., <maths id="math0016" num="(12)"><math display="block"><msubsup><mi>P</mi><mi>b</mi><mi mathvariant="italic">peak</mi></msubsup><mo>=</mo><mi>diag</mi><mfenced open="{" close="}"><mi mathvariant="italic">abs</mi><mfenced><msub><mi>W</mi><mi mathvariant="italic">tx</mi></msub></mfenced><mo>⁢</mo><mi mathvariant="bold">I</mi><mspace width="1em" /><mi mathvariant="italic">abs</mi><mfenced><msubsup><mi>W</mi><mi mathvariant="italic">tx</mi><mo>*</mo></msubsup></mfenced></mfenced><mo>,</mo></math><img file="EP0804858B1_D0016.tif" /></maths> where I is a matrix of all ones of the appropriate size and abs(.) is elementwise absolute value. Channel selector 35 compares these values against the limits for each of the transmitters for each of the elements. If any of the average or peak values exceed the acceptable limits, the remote terminal in question is not assigned to the candidate channel. Otherwise, the ability to successfully receive from the remote terminal is checked. In an alternate embodiment, the transmitter limits are used as inequality constraints in an optimization algorithm for calculating transmit weights that meet the specifications given and that also result in the minimum amount of transmitted power possible. If transmit weights satisfying the constraints can not be found, the remote terminal in question is not assigned to the candidate channel. Such optimization algorithms are well known.

0068To test the uplink, channel selector 35 calculates spatial demultiplexing weights <i>W_rx</i> using <i>A_br,</i> the demultiplexing spatial signature matrix <i>A_br.p</i> associated with the relevant conventional channel, as given for the preferred embodiment in equation (13): <maths id="math0017" num="(13)"><math display="block"><msub><mi>W</mi><mi mathvariant="italic">rx</mi></msub><mo>=</mo><msup><mfenced><msub><mi>A</mi><mi mathvariant="italic">br</mi></msub><mo>⁢</mo><msub><mi>P</mi><mi>r</mi></msub><mo>⁢</mo><msubsup><mi>A</mi><mi mathvariant="italic">br</mi><mo>*</mo></msubsup><mo>+</mo><msub><mi>R</mi><mi mathvariant="italic">nn</mi></msub></mfenced><mrow><mo>-</mo><mn>1</mn></mrow></msup><mo>⁢</mo><msub><mi>A</mi><mi mathvariant="italic">br</mi></msub><mo>⁢</mo><msub><mi>P</mi><mi>r</mi></msub><mo>,</mo></math><img file="EP0804858B1_D0017.tif" /></maths> where <i>P_r</i> is a (diagonal) matrix of mean-square amplitudes (powers) transmitted by the remote terminals and <i>R_nn</i> is the base station receiver noise covariance. Then, the expected value of the normalized mean-squared error covariance is calculated in one embodiment as follows: <maths id="math0018" num="(14)"><math display="block"><mover><mi mathvariant="italic">MSE</mi><mo>‾</mo></mover><mo>=</mo><msubsup><mi>P</mi><mi>r</mi><mrow><mo>-</mo><mn>1</mn><mo>/</mo><mn>2</mn></mrow></msubsup><mo>⁢</mo><mfenced><mfenced><mi>I</mi><mo>-</mo><msubsup><mi>W</mi><mi mathvariant="italic">rx</mi><mo>*</mo></msubsup><mo>⁢</mo><msub><mi>A</mi><mi mathvariant="italic">br</mi></msub></mfenced><mo>⁢</mo><msub><mi>P</mi><mi>r</mi></msub><mo>⁢</mo><msup><mfenced><mi>I</mi><mo>-</mo><msubsup><mi>W</mi><mi mathvariant="italic">rx</mi><mo>*</mo></msubsup><mo>⁢</mo><msub><mi>A</mi><mi mathvariant="italic">br</mi></msub></mfenced><mo>*</mo></msup><mo>+</mo><msubsup><mi>W</mi><mi mathvariant="italic">rx</mi><mo>*</mo></msubsup><mo>⁢</mo><msub><mi>R</mi><mi mathvariant="italic">nn</mi></msub><mo>⁢</mo><msub><mi>W</mi><mi mathvariant="italic">rx</mi></msub></mfenced><mo>⁢</mo><msubsup><mi>P</mi><mi>r</mi><mrow><mo>-</mo><mo>*</mo><mo>/</mo><mn>2</mn></mrow></msubsup></math><img file="EP0804858B1_D0018.tif" /></maths> The notation (·)^-*/2 indicates complex conjugate transpose of the square root of the matrix. The inverse of <i>MSE</i> is an estimate of the expected Signal-to-interference-plus-Noise Ratio (<i>SINR</i>) at the output of the spatial demultiplexer: <maths id="math0019" num="(15)"><math display="block"><mover><mi mathvariant="italic">SINR</mi><mo>‾</mo></mover><mo>=</mo><msup><mover><mi mathvariant="italic">MSE</mi><mo>‾</mo></mover><mn>1</mn></msup><mn>.</mn></math><img file="EP0804858B1_D0019.tif" /></maths> If all of the diagonal elements of <i><o ostyle="single">SINR</o></i> are above the desired thresholds based on the signal quality required to be received from each remote terminal, the remote terminal is allowed access to the channel. If the candidate remote terminal is below its threshold and has the ability to increase its output power, the same computations are again performed for increasing remote terminal power output until either the maximum output power for that remote terminal is reached and the <i><o ostyle="single">SINR</o></i> is still insufficient, another remote terminal <i><o ostyle="single">SINR</o></i> falls below its threshold in which case its power is increased if possible, or all thresholds are exceeded. If acceptable remote terminal transmit powers can be found, the remote terminal is granted access to this particular conventional channel, otherwise it is denied access and another conventional channel is checked.

0069In an alternate embodiment, the calculation of demultiplexing weights is performed using well known optimization procedures with the objective of minimizing remote terminal transmit powers subject to estimated signals at the base station meeting or exceeding their minimum desired <i>SINR's.</i>

0070Also, in an alternate embodiment, in the case that no conventional channel can be found to accommodate the remote terminal, channel selector 35 calculates whether some rearrangement of the existing remote terminals among the conventional channels would allow the remote terminal to be supported on some conventional channel. In this case, the remote terminal will only be denied communication at this time if no rearrangement of existing users allows the remote terminal to be accommodated.

0071In an alternate embodiment employing frequency division duplexing (FDD), remote terminals are not restricted to being assigned a fixed conventional channel pair for transmit and receive. A sufficiently flexible system architecture is employed where channel selector 35 may choose to assign a particular remote to transmit and receive conventional channels separated by different frequency duplex offsets in order to minimize overall system interference levels.

0072Spatial multiplexing and demultiplexing weights for remote terminals already using a conventional channel must be recalculated because adding a new remote terminal to that conventional channel may change them significantly. In the preferred embodiment, channel selector 35. having already done the necessary calculations, sends the new spatial multiplexing and demultiplexing weights to the spatial weight processor 37 for use in setting up the spatial multiplexer 23 and demultiplexer 20. In an alternate embodiment, spatial weight processor 37 uses the spatial signature matrices sent to it by channel selector 35 to calculate different sets of spatial multiplexing and demultiplexing weights for all of the remote terminals on that conventional channel.

0073Spatial weight processor 37 then sends the new spatial demultiplexing weights to spatial demultiplexers 20 and the new spatial multiplexing weights to the spatial multiplexers 23 for this conventional channel, updates the active remote terminal list 34, and informs spatial processor controller 33 which in turn informs base station controller 3 of the selected channel. Base station controller 3 then transmits a message to the remote terminal using the downlink control channel that instructs the remote terminal to switch to the desired conventional channel.

0074It can be shown from equation (9) that the multiplexing weight matrices <i>W_tx</i> have the property: <maths id="math0020" num="(16)"><math display="block"><msub><mi>A</mi><mi mathvariant="italic">rb</mi></msub><mo>⁢</mo><msubsup><mi>W</mi><mi mathvariant="italic">tx</mi><mo>*</mo></msubsup><mo>=</mo><msub><mi>S</mi><mi>b</mi></msub><mn>.</mn></math><img file="EP0804858B1_D0020.tif" /></maths> This means that at the <i>i^th</i> remote terminal, the signal intended to be sent to that terminal is received with a sufficient (positive real) amplitude <i>S_b</i>(<i>i, i</i>). The fact that <i>S_b</i> has zero off-diagonal elements means that at the <i>i^th</i> remote terminal, none of the other signals being transmitted are received by that remote terminal. In this manner, each remote terminal receives only the signals intended for it at the necessary power levels to ensure proper communications. In alternate embodiments, uncertainties in the estimates of <i>A_rb</i> are incorporated in setting base station transmit power levels and calculating weights so as to minimize the effect of errors and/or changes in <i>A_rb.</i>

0075Similarly, at the base station the particular demultiplexing weight matrices given in (13) have the property that conditioned on the knowledge of the receive spatial signatures and the transmitted voltages (powers) from the remote terminals, the estimated signals <i>Ŝ</i> given by: <maths id="math0021" num="(17)"><math display="block"><mover><mi>S</mi><mo>^</mo></mover><mo>=</mo><msubsup><mi>W</mi><mi mathvariant="italic">rx</mi><mo>*</mo></msubsup><mo>⁢</mo><msub><mi mathvariant="normal">z</mi><mi>b</mi></msub><mo>,</mo></math><img file="EP0804858B1_D0021.tif" /></maths> are the most accurate in the sense of least mean-squared error. In particular, they most closely match the signals transmitted by the remote terminals given the measurements made at the base station by the multiple antenna elements.

0076Equations (9) and (13) represent only one way to calculate spatial multiplexing and demultiplexing weights. There are other similar strategies that demonstrate properties similar to those shown in equation (16) and described in the previous paragraph. Other well known techniques for calculating weight matrices <i>W_tx</i> and <i>W_rx</i> account for uncertainty in spatial signature matrices <i>A_rb</i> and <i>A_br</i>, for wide bandwidth conventional channels, and can incorporate more complex power and dynamic range constraints.

Determining Spatial Signatures

0077As shown in <figref idref="f0006">figure 6</figref>. spatial processor 13 also contains a spatial signature processor 38 for finding the spatial signatures of the remote terminals. In the illustrative embodiment, spatial signature processor 38 uses the calibration techniques described in copending <patcit id="pcit0002" dnum="US234747A" dnum-type="L"><text>U.S. patent application 08/234,747</text></patcit>.

0078In the illustrative embodiment, each remote terminal is capable of entering a calibration mode in which the signal received by the remote terminal 43 is transmitted back to base station 1. Referring to <figref idref="f0007">figure 7</figref>, this function is provided by switch 63 controlled by remote terminal CPU 62 through switch control signal 64.

0079To determine the transmit and receive spatial signatures of a remote terminal, spatial signature processor 38 commands the remote terminal to enter calibration mode by transmitting a command to it on the downlink channel. This command is generated by base station controller 3. based on a request from spatial processor controller 33. and modulated by signal modulators 24. Spatial signature processor 38 then transmits predetermined calibration signals 11, on the conventional channel occupied by the remote terminal, by instructing multichannel transmitters 17(a, ....m) via transmitter control data 31 and spatial processor controller 33. In the present embodiment, the m signals (for each antenna) among the predetermined calibration signals 11 are different frequency complex sinusoids. In another embodiment, the predetermined calibration signals 11 are any known, distinct, signals.

0080The remote terminal shown in <figref idref="f0007">figure 7</figref> transmits back the signal received at the remote terminal. This transponded signal is received by multichannel receivers 15 in base station 1 shown in <figref idref="f0001">figure 1</figref> and supplied to spatial signature processor 38 shown in <figref idref="f0006">figure 6</figref>. In one embodiment described in patent application 08/234,747. spatial signature processor 38 computes the receive and transmit spatial signatures of the remote terminal from the received signal measurements 6 and predetermined calibration signals 11 as follows. Time samples of the received data are stored in an m by n data matrix Z which in the absence of noise and parameter offsets is given by <maths id="math0022" num="(18)"><math display="block"><mi mathvariant="normal">Z</mi><mo>=</mo><msub><mrow><msub><mi mathvariant="italic">ka</mi><mi mathvariant="italic">br</mi></msub><mo>⁢</mo><mi mathvariant="normal">a</mi></mrow><mi mathvariant="italic">rb</mi></msub><mo>⁢</mo><mi mathvariant="normal">S</mi><mo>,</mo></math><img file="EP0804858B1_D0022.tif" /></maths> where S is the m by n matrix of predetermined calibration signals and k is a known amount by which the signal is amplified in the remote terminal before transmission back to the base station. The receive spatial signature is proportional to the singular vector (u₁) corresponding to the largest singular value (σ<i>_max</i> of the data matrix Z. Transmission of a unit power signal from the remote terminal and received by the base station at antenna element 1 provides the necessary scaling <i>g_br</i> for the receive spatial signature <maths id="math0023" num="(19)"><math display="block"><msub><mi mathvariant="normal">a</mi><mi mathvariant="italic">br</mi></msub><mo>=</mo><msub><mi>g</mi><mi mathvariant="italic">br</mi></msub><mo>⁢</mo><msub><mi mathvariant="normal">u</mi><mn mathvariant="normal">1</mn></msub><mo mathvariant="normal">/</mo><msub><mi mathvariant="normal">u</mi><mn mathvariant="normal">1</mn></msub><mfenced><mn>1</mn></mfenced><mn>.</mn></math><img file="EP0804858B1_D0023.tif" /></maths> where u₁(1) is the first element of u₁. Once a<i>_br</i> is known, a<i>_rb</i> is calculated by <maths id="math0024" num="(20)"><math display="block"><msub><mi mathvariant="normal">a</mi><mi mathvariant="italic">rb</mi></msub><mo>=</mo><msup><mi>k</mi><mrow><mo>-</mo><mn>1</mn></mrow></msup><mo>⁢</mo><msup><mfenced><msub><mi>g</mi><mi mathvariant="italic">br</mi></msub><mo>⁢</mo><msub><mi mathvariant="normal">u</mi><mn mathvariant="normal">1</mn></msub><mo mathvariant="normal">/</mo><msub><mi mathvariant="normal">u</mi><mn mathvariant="normal">1</mn></msub><mfenced><mn>1</mn></mfenced></mfenced><mi>†</mi></msup><mo>⁢</mo><msup><mi>zs</mi><mi>†</mi></msup><mn>.</mn></math><img file="EP0804858B1_D0024.tif" /></maths> where <i>B</i>^† is the well known Moore-Penrose pseudo-inverse of the matrix <i>B</i> satisfying <i>BB</i>^† = <i>I (the</i> identity matrix) for full-rank matrices <i>B</i> having more columns than rows. <i>B</i>^†<i>B</i> = I for full-rank matrices <i>B</i> having more rows than columns. In alternate embodiments also described in copending application 08/234.747. well known techniques are used to account for noise present in the system and parameter variations such as oscillator frequency offsets.

0081Spatial signature processor 38 stores the new spatial signatures in remote terminal database 36. Upon completion, spatial signature processor 38 commands the remote terminal to exit calibration mode by transmitting a command to it on the downlink channel.

0082In one alternate embodiment, computation of remote terminal transmit spatial signatures can be performed directly by the remote terminals. This embodiment of the remote terminal is shown in <figref idref="f0008">figure 8</figref>. In calibration mode, spatial signature processor 38 transmits predetermined calibration signals 11. on the conventional channel to be calibrated by the remote terminals, as before. Remote terminal CPU 62 uses received calibration signals 44 and the known transmitted waveforms to compute the remote terminal's transmit spatial signature using the same techniques used by spatial signature processor 38 in the previous embodiment. The computed transmit spatial signature is transmitted back to base station 1 through modulator 51 and transmitter 54 as control data to be transmitted 52. When received by base station 1. spatial signature processor 38 stores the new transmit spatial signature in remote terminal database 36. Since each remote terminal performs the transmit spatial signature calculation independently, this arrangement allows multiple remote terminals to compute their own transmit spatial signature simultaneously on the same conventional channel. In this embodiment, remote terminal receive spatial signatures are computed by spatial signature processor 38 in the same manner as in the previous embodiment.

0083Using these techniques, spatial signature processor 38 can measure a remote terminal transmit and receive spatial signatures for a particular channel any time that channel is idle. The efficiency of these calibration techniques allow spatial signature processor 38 to update the spatial signatures of numerous remote terminals for a particular channel while occupying that channel for only a short time.

0084Many other techniques for obtaining spatial signatures of remote terminals are also available. In some RF environments, spatial signatures for remote terminals can be determined using well-known techniques that depend upon knowledge of the geometric arrangement of the m reception antennas 19(a.....m) and their individual directivity patterns (element gain and phase, with respect to a reference, as a function of angle-of-arrival), and the direction from the base station to the remote terminal. Furthermore, techniques such as ESPRIT (<patcit id="pcit0003" dnum="US4750147A"><text>U.S. patents 4.750.147</text></patcit> and <patcit id="pcit0004" dnum="US4965732A"><text>4.965.732</text></patcit>) can be used to estimate directions in applications where they are not known a priori.

0085Similarly, as is well known, knowledge of any predetermined modulation format parameters of the underlying signals being transmitted by the remote terminals (for example, knowledge of certain training or preamble sequences, or knowledge that the signals are constant modulus) can also be used to determine the receive spatial signatures for remote terminals. A further example is decision-directed feedback techniques, also well known in the art, where receive data is demodulated and then remodulated to produce an estimate of the original modulated signal. These techniques allow receive spatial signatures to be estimated even when multiple remote terminals are occupying a single conventional channel.

0086In some RF environments, transmit spatial signatures for remote terminals can be calculated explicitly, as is well known, using knowledge of the remote terminal locations and the locations and directivity patterns of the base station transmit antennas. This requires no special capability on the part of the remote terminal.

0087If the remote terminal has the ability to measure and report the strength of the signal it is receiving, the system can use this information to derive transmit spatial signatures, albeit in a less efficient manner than the embodiment shown in <figref idref="f0007">figure 7</figref> where the remote terminal has full transponder capabilities, or the embodiment shown in <figref idref="f0008">figure 8</figref> where the remote terminal directly computes its transmit spatial signature. The transmit spatial signature is determined based solely on received signal power reports from the remote terminal as follows. First, spatial signature processor 38 transmits identical unit power signals from two of the m antenna elements at a time. Spatial signature processor 38 then changes the amplitude and phase of one of the two signals until the remote terminal reports that it is receiving no signal. The set of complex weights for antenna elements 2 through m required to null a unit power signal from element 1 are changed in sign and inverted to produce the transmit spatial signature for the remote terminal.

0088In yet another embodiment, the system can be designed to continuously update the spatial signatures of the remote terminals in a "closed loop manner. This is done to account for the time variation of spatial signatures due to, for example, motion of the remote terminal or changes in the RF propagation conditions. To do this, both the base station and the remote terminal periodically transmit predetermined training sequences. Each remote terminal currently active on a particular channel is assigned a different predetermined training sequence and is given the training sequences for all other remotes currently active on that particular channel. In one embodiment, the different training sequences are orthogonal in the sense that the inner product of any two of the training sequence waveforms is zero. Each time the training sequences are transmitted, each remote terminal calculates how much of each training sequence it has received using well known techniques, and transmits this information to the base station.

0089In the illustrative embodiment, the base station uses the receiver outputs and knowledge of the transmitted waveforms to calculate the remote terminal receive spatial signatures. In another embodiment, the base station calculates how much of each remotely transmitted training sequence has come through on each output of the spatial demultiplexer, expressed as a complex vector of coupling coefficients. Knowledge of these coupling coefficients allows the currently active receive and transmit spatial signatures to be corrected so as to reduce mutual interference using well know techniques.

0090Finally, in systems that use time division duplexing (TDD) for full-duplex communications, as is well known in the art, the transmit and receive frequencies are the same. In this case, using the well known principle of reciprocity, the transmit and receive spatial signatures are directly related. Thus, this embodiment determines only one of the signatures, for example the receive spatial signature, and the other, in this case the transmit spatial signature, is calculated from the first (receive) spatial signature and knowledge of the relative phase and amplitude characteristics of multichannel receivers 15 and multichannel transmitters 14.

Network Level Spatial Processing

0091In the embodiment illustrated herein, the spatial processor for each base station in the cellular-like wireless communication system operates independently to maximize the number of communication channels in the immediate cell. However, significant system capacity improvements can be realized if the spatial processor from each base station communicates with and coordinates its efforts with the spatial processors from other nearby cells. A specific embodiment is shown in <figref idref="f0009">figure 9</figref>.

0092A multiple base station controller 66 acts as the interface between the wide area network 65 through link 68 and base stations 1 (a,b,c) via base station communication links 2 (a.b.c). Each base station is responsible for providing coverage to a number of remote terminals. In one embodiment, each remote terminal is assigned to only one base station thus defining cell boundaries 67 (a.b.c) within which all remotes attached to a particular base station are located. Users equipped with remote terminals 69 are identified by a boxed "R" in the figure.

0093Each spatial processor contained in base stations 1 (a.b,c) measures and stores the spatial signatures of the remote terminals in its cell and also of the remote terminals in adjacent cells. The determination of spatial signatures of the remote terminals in adjacent cells is coordinated by multiple base station controller 66 through base station communication links 2 (a.b.c). Through base station communication links 2 (a.b.c) and multiple base station controller 66, spatial processors in base stations 1 (a,b.c) from adjacent cells inform each other of which remote terminals they are communicating with on which conventional channels. Each spatial processor includes the spatial signatures of remote terminals that are currently active in adjacent cells to form extended transmit and receive spatial signature matrices <i>A_rb</i> and <i>A_br</i> which are sent to all the adjacent base stations. The channel selectors in each base station, using these extended spatial signature matrices, jointly assign remote terminals to each conventional channel in each of base stations 1 (a,b,c).

0094The resultant weight matrices <i>W_tx</i> and <i>W_rx</i> for each base station are then calculated using extended spatial signature matrices <i>A_rb</i> and <i>A_br.</i> In calculating the weights, the objective is to minimize the signal transmitted to and received from the adjacent cell's active remote terminals, thereby allowing many more remote terminals to simultaneously communicate.

0095In an alternate embodiment, multiple base station controller 66 assigns remote terminals requesting access to base stations dynamically using a list of active remote terminal/base station/conventional channel links, the associated remote terminal databases, and the particular requirements for the link to be assigned. Additionally, remote terminals can employ multiple (directional) transmit and receive antennas, to facilitate directive links to multiple nearby base stations as instructed by multiple base station controller 66 to further increase system capacity.

<i>Advantages</i>

0096The apparatus and method m accordance with the invention provides a significant advantage over the prior art in that it allows many remote terminals to simultaneously share the same conventional communication channel. In particular, for a system with m receive and m transmit antenna elements, up to m remote terminals can share a single conventional communication channel. Moreover, signals received from and transmitted to the remote terminals have substantially improved signal-to-noise, reduced interference, and improved quality in multipath environments compared to a standard base station.

0097Thus, a wireless communication system can support many times more conversations, or have a much greater data throughput, with the same amount of spectrum. Alternatively, a wireless communication system can support the same number of conversations or data throughput with much less spectrum.

<i>Alternate Embodiments</i>

0098In one alternate embodiment, transmission antennas 18(a,...,m) and reception antennas 19(a,....m) at base station 1 are replaced by a single array of m antennas. Each element in this array is attached to both its respective component of multichannel transmitters 14 and its respective component of multichannel receivers 15 by means of a duplexer.

0099In another alternate embodiment, signals on the uplink control channel may be processed in real time using the spatial processing described in copending patent application 07/806.695. This would allow multiple remote terminals to request a communication channel at the same time.

0100In yet another embodiment for applications involving data transfer of short bursts or packets of data, no separate uplink control channel is required and the system may service requests for communication and other control functions during control time intervals that are interspersed with communications intervals.

0101As stated above, many techniques are known for measuring the spatial signatures of the remote terminal radios and using these spatial signatures to calculate multiplexing and demultiplexing weights that will allow multiple simultaneous conversations and/or data transfers on the same conventional communication channel.

0102While the above description contains many specificities, these should not be construed as limitations on the scope of the invention, but rather as an exemplification of one preferred embodiment thereof. Many other variations are possible. Accordingly, the scope of the invention should be determined not by the illustrated embodiments, but by the appended claims and their legal equivalents.