A method for generating a dynamic and / or random voxelized environment model for signal processing in wireless communication and / or environment perception, and a device for implementing the method and / or applying the output thereof.

Abstract

A method for processing a voxelized environment model (VEM) of a region of interest (ROI) for use in wireless communication signal processing and / or environment recognition or mapping comprises the step of obtaining one or more first and second representations of the VEM of the ROI determined from or based on one or more first and second sets of transmitted symbols received at one or more corresponding first and second times by one or more transmitters to or across the ROI and at one or more receivers. The first and second representations of the VEM of the ROI are compared with regard to each first and second time to extract or derive auxiliary voxel information (AVI), and a predicted VEM of the ROI is generated based on the representation of the VEM of the ROI and the AVI. The predicted VEM of the ROI is output for further use.

Description

Technology Field

[0001] The present invention relates to the fields of wireless communication and environmental perception, and more specifically, to wireless communication that generates a spatial representation of an environment and / or optimizes the operating parameters of a transmitter and / or receiver using information related to the spatial representation of an environment. Generating a spatial representation of an environment is also referred to as environmental perception, and generating a spatial representation of an environment using wireless communication signals is also referred to as Joint Communication and Sensing Integration (JCAS).

[0002] Throughout this specification, the terms environmental perception, environmental mapping, or environmental sensing may be used to refer to various representations widely used to acquire information about the environment and generate a three-dimensional representation. Background Technology

[0003] Wireless communication can be affected by objects located between the transmitter and the receiver; these objects can, in particular, block or attenuate the signal, cause reflections that lead to multipath reception, or cause phase shifts at the receiver. Advanced wireless receivers typically attempt to estimate a channel matrix composed of elements representing the characteristics of the wireless communication channel, such as attenuation, gain, and phase shift, within the frequency range used for communication. This estimated channel matrix is ​​used to equalize the received signal prior to symbol detection. Obtaining appropriate information about the environment—specifically the objects located between the transmitter and the receiver—facilitates the estimation of the channel matrix and / or improves its accuracy, which can reduce the computational complexity required for estimation and / or symbol detection. In addition to reducing the electrical energy required for computation, reducing computational complexity can shorten the time between the time a signal is received at the antenna and the time the transmitted symbol becomes available at the receiver's output.

[0004] Information about the environment between the transmitter and the receiver can also be utilized for other purposes, particularly in generating environment maps that can be used for various applications, such as pathfinding through the environment and collision avoidance, especially on mobile devices.

[0005] For example, various technologies are known for generating environmental representations using cameras, radar, lidar, or combinations thereof, but many of these known technologies require additional equipment and do not necessarily provide information or formats favorable for wireless communication.

[0006] Spatial representations of environments are widely used in robot vision, localization, and mapping, as well as in computer graphics and medical imaging. One known technique involves discretely approximating the real environment using a voxelized occupancy grid. Typically, a voxel is a cube arranged in a 3D space covering the environment, and information about the space occupied by the voxel is assigned to the voxel. This discreteness of the voxelized space is highly suitable for 3D modeling, and objects in the region of interest can be represented as approximate estimates or in a more accurate form depending on the size of the unit voxel.

[0007] An exemplary environment is illustrated in FIG. 1a). The entire region of interest (ROI) is defined such that each represents the lengths of the x, y, and z axes in meters. L x × L y × L z It is defined as a cubic space of size. The entire ROI is It is subdivided into a grid consisting of voxels, where and represents the number of voxels in the x, y, and z axis directions, respectively, and L V This represents the length of the edges of a voxel cube in meters and corresponds to image resolution. 3D tensor( N x × N y × N z When expressed as ), the voxelized occupancy grid directly represents a discretized model of the ROI as illustrated in Fig. 1 b) and c), where the size of the voxel corresponds to the image resolution. The elements of the 3D tensor represent the occupancy state of the voxel, that is, whether the space is empty or filled with a specific material. In addition to the 3D geometric information provided by the classical voxelized occupancy grid, modeling methods suitable for wireless communication scenarios can be developed, including the electromagnetic scattering behavior of the real environment along with any additional data or information required for each application or use case. In this specification, the voxelized occupancy grid is also referred to as the voxelized environment model.

[0008] More recently, Joint Communication and Sensing (JCA) technology has been developed, a wireless communication technology aimed at achieving data communication while simultaneously extracting environmental information from signal scattering present in the Effective Channel State Information (CSI) caused by, for example, objects within the environment, shielding, and user activity. Most known JCA methods infer environmental information by utilizing radar technology. This is also known as Joint Radar and Communication (JRC).

[0009] Various methods are known in JRC, including spectrum alternation or sharing between radar and communication signals, information embedding using standard radar signals, radar parameter extraction from standard communication signals, or even new waveform designs suitable for both purposes. These known techniques are heavily based on conventional radar signal processing, for example, ambiguity function estimation, rely on the frequency-delay characteristics of radar, and suffer from similar problems.

[0010] Recent advancements in communication technology have revealed that modeling the space between a transmitter and a receiver as a voxelized environment is beneficial for JCAS, often referred to as sensing and communication integration technology or ISAC.

[0011] To further enhance the benefits of environmental representation for communication purposes, communication channel modeling methods used in wireless communication scenarios can be optimized by adding the electromagnetic scattering behavior of objects in the real environment to the 3D geometric information provided by classical voxelized occupancy grids. To account for scattering behavior, each voxel has a voxel occupancy coefficient v k ∈ {0, 1}(here, k ∈ {1, ..., N V }) is assigned, and here v k = 0 indicates that the k-th voxel is empty, meaning the corresponding environment is free space, and v k = 1 indicates that the k-th voxel is occupied by a scatterer (e.g., an object on a table, chair, or wall) (see Fig. 1). The binary voxel occupancy coefficient is used to capture the effect occurring on electromagnetic waves reflected by the occupied voxel. complex Voxel scattering coefficient, that is It should be noted that it can be extended to. Constant β k and ω k It depends not only on the material itself but also on the frequency and angle of incidence of the radio signal, and represents the effect that the material occupying a given voxel has on the electromagnetic waves reflected or refracted by this voxel. In other words, the voxel scattering coefficient value is expected to depend significantly on the electromagnetic properties of the scatterer and the incident wave, and these properties can be empirically measured and modeled as a function of properties such as frequency and material.

[0012] For example, the literature ["Joint Multi-User Communication and Sensing Exploiting Both Signal and Environment Sparsity," IEEE Journal of Selected Topics in Signal Processing, vol. 15, no. 6, pp. 1409-1422, Nov. 2021, X. Tong, Z. Zhang, J. Wang, C. Huang and M. Debbah] considers a regular voxelized 3D space with multiple scatterers accommodating a single access point (AP), a single reconfigurable intelligent surface (RIS), and multiple single-antenna user devices (UEs). The multiple UEs communicate with the AP via a sparse code multiple access (SCMA) method, through line-of-sight (LOS) and non-line-of-sight (NLOS) paths, from the UE through the scatterers and RIS to finally reach the AP, using multiple frequency subcarriers and multiple transmit instances.

[0013] Figure 2 illustrates a schematic representation of the 3D space considered in known systems and methods, including the RIS. Here, signals reflected from the RIS toward the AP are indicated by dashed lines to highlight specific starting points. This figure shows direct or LOS signal paths and indirect or NLOS signal paths.

[0014] Figure 3 illustrates the general concepts of LOS and NLOS paths in a voxelized space where all paths are available and the reflection angle is within the range that reflects the incident electromagnetic waves. In this figure, the LOS path is a direct path between a User Equipment (UE) and an Access Point (AP), whereas two occupied voxels within the ROI reflect signals emitted from the UE toward the AP. The dashed line represents the NLOS UE-voxel path, and the dotted line represents the NLOS voxel-AP path.

[0015] In addition to the flexibility of estimation accuracy provided by voxelized grids, one of the biggest advantages of grid-based models is the simplicity of data representation, particularly compared to 3D vertex-based or point cloud-based methods, which is more advantageous from a machine learning perspective.

[0016] Like most, but not all, other known methods for implementing JCAS in a voxelized environment, the prior art method described above relies on ideal assumptions that all paths between all transmitting and receiving antennas are fully available, i.e., there are no blocked paths, that the channel gain of the UE-AP LOS path is known, that the channel gains of the UE-voxel, voxel-RIS, and RIS-AP NLOS paths are known, that the RIS reflection coefficient is known, that the voxelized environment model is a binary model, i.e., only discrete occupancy values ​​of 0 or 1 are possible, that all scattering paths have realistic reflection angles, that the SCMA communication scheme is used, and that only a single AP exists. However, some paths may not be available for radio signals, while others may be available.

[0017] Figure 4 illustrates the general concept of impossible paths, or unavailable paths, between two UEs and an AP. The path between UE1 and the AP has a reflection angle so shallow that a signal cannot actually propagate from an occupied voxel to the AP. The path between UE2 and the AP is the most direct path, but it is blocked by an occupied voxel. It should be noted that no possible paths are illustrated in this figure.

[0018] Furthermore, known methods do not consider energy loss during reflection scattering that varies depending on the material and surface structure of the scatterer, frequency-dependent scattering behavior, as well as the relative distance between the voxel and the device and the resulting effect on channel gain.

[0019] Finally, although known methods may be able to provide a sufficiently accurate voxelized representation of the environment, they consider only static environments—that is, environments where the positions of occupied voxels do not change—and therefore do not consider the time required to obtain a somewhat complete voxelized representation of the environment, or even if considered, it is not regarded as important.

[0020] As such, the idealized assumptions found in conventional methods cause serious limitations in applying known methods to actual 3D real-world environments and applications, thereby creating a need for an improved method for generating a voxelized environment model for signal processing in wireless communication, a transmission and reception method using such an improved voxelized environment model, and a device for implementing such a method.

[0021] These requirements are addressed by the methods of claims 1, 7, and 9, the apparatus of claims 6 and 10, and the computer program product of claim 11. Claim 12 provides a corresponding computer-readable storage medium. Each dependent claim provides embodiments and improvements.

[0022] According to a first aspect of the present invention, a method for processing a voxelized environment model of a region of interest (ROI) for use in wireless communication signal processing and / or environment recognition or mapping is provided.

[0023] The method includes the step of obtaining one or more first representations of a voxelized environment model (VEM) of an ROI determined from or based on one or more first transmission symbol sets received at one or more corresponding first time points by one or more transmitters to or across an ROI and one or more receivers.

[0024] The method further comprises the step of one or more transmitters transmitting wirelessly to or across the ROI, and one or more receivers obtaining one or more second representations of the ROI's VEM determined from or based on one or more second transmission symbol sets received at one or more corresponding second timestamps different from each first timetamp. The different first and second timestamps may be separated by one or more transmission intervals, for example, as defined in wireless communication standards, etc. It is assumed that wireless communication standards allow the different transmission intervals to be clearly identified, for example, through timestamps, fixed transmission intervals, etc. For clarity, it is assumed that the second timetamp occurs after the first timetamp. Inverting the first and second timestamps is simply accomplished by inverting the corresponding input values ​​for each method step, and it is evident that such an invention is also within the scope of the present invention.

[0025] The step of obtaining first and second representations of the VEM of the ROI may include receiving a corresponding first and / or second set of transmitted symbols and deriving each representation of the VEM of the ROI therefrom. The deriving step may include an iterative or non-iterative method comprising the process disclosed in German patent application number DE 10 2022 212 615 A1 (the whole of which is incorporated herein by reference).

[0026] The transmitted symbol may be received at first and second timestamps at one or more antennas each connected to the device executing the method. In this context, the expression 'connected' may refer to a direct electrical connection or any other connection that provides the received signal or an appropriate representation thereof to the device executing the method.

[0027] Transmission symbols transmitted by one or more transmitters may be received at a single receiver or at multiple receivers. Each of the one or more receivers may determine the VEM of the ROI from each of the respective received transmission symbols, and one or more resulting VEMs of the ROI determined by each of the one or more receivers, or a processed combination thereof, e.g., a merged representation, etc., may be used in a subsequent step of the method.

[0028] The first and / or second representation of the VEM of the ROI may include at least one occupancy value for each voxel, which indicates whether the voxel is empty or occupied. An empty voxel may be considered not to contain any object or material that significantly affects the propagation of radio waves incident on the voxel. Within the Earth's atmosphere, an empty voxel may contain not only gaseous elements constituting the atmosphere but also aerosols or very small solid particles suspended in the atmosphere that cause some attenuation of radio waves but do not significantly affect the propagation of radio waves. Significant effect on the propagation of radio waves may mean strong attenuation, blocking or reflection of radio waves, etc., which do not generally occur in open space. The occupancy value may additionally include or be supplemented by a value indicating the probability that the occupancy value is true or accurate. In an embodiment of the method, the occupancy value may additionally include or be supplemented by a value representing a critical angle at which the signal is no longer reflected and is blocked, a value representing signal attenuation and / or phase shift at reflection and dependence on the angle of incidence, attenuation according to frequency, reflection and / or blocking, etc.

[0029] In the next step, the method includes the step of comparing one or more first representations of the VEM of the ROI and one or more second representations of the VEM of the ROI, taking into account each first and second time point, in order to extract or derive auxiliary voxel information (AVI).

[0030] In the context of this description, the term "comparison" may include applying advanced analytical methods, such as machine learning (ML) or artificial intelligence (AI) tools, to extract or derive AVI, in addition to simple value-for-value comparison. Such advanced analytical methods may include, but are not limited to, spatial target tracking methods for identifying voxels associated with stationary and moving objects, respectively.

[0031] The method further includes the step of generating a predicted VEM of the ROI at, for example, a current or future time based on one or more second representations of the voxelized environment model of the region of interest and auxiliary voxel information extracted / derived in the comparison step. The predicted VEM of the ROI includes a predicted occupancy value for each voxel. The predicted occupancy value may be supplemented with each value representing the probability that the predicted occupancy value is true or accurate. Finally, the predicted VEM of the ROI is output to, for example, a process for detecting transmitted symbols, a process for determining the environment model of the region of interest based on received transmitted symbols, and / or a sensor fusion process for determining the environment model of the region of interest.

[0032] When output to the process of detecting transmitted symbols at the receiver, the predicted voxelized environment can particularly improve relevant channel estimation and signal equalization.

[0033] Similarly, when a predicted voxelized environment is output to a process that determines an environment model of a region of interest based on received transmitted symbols, the accuracy of the determined environment model may be improved and / or inevitable errors related to the prediction may be reduced.

[0034] Similarly, when the predicted voxelized environment is output to a sensor fusion process that determines the environment model of the region of interest, the corresponding environment model can benefit from additional inputs.

[0035] It is evident that the predicted VEM can be used as a first representation of the VEM for the next execution of the method and / or as a preliminary or preparatory input for each process that determines one or more second representations of the VEM to be input to the method.

[0036] As mentioned above, the extracted or derived AVI may include information regarding changes in the position or spatial location of one or more occupied voxels between the first and second representations of the VEM of the ROI in one or more embodiments, e.g., position, orientation, and / or velocity information. Changes in the position of one or more voxels may be determined only for voxels where a value representing the probability of occupancy and / or the probability that the predicted occupancy value is true or accurate exceeds a respective predetermined threshold. Changes in the position or spatial location may be determined for one or more antennas where the signal is processed in the device executing the method.

[0037] One or more voxels, preferably adjacent voxels, may be grouped into one or more clusters, each cluster preferably containing coherent voxels, and positional changes may be determined for the entire cluster. Grouping may be performed on voxels where the occupancy probabilities are similar and / or where the value representing the probability that the predicted occupancy value is true or accurate exceeds a predetermined threshold. This grouping can reduce the amount of computation.

[0038] Information on changes in the location of one or more occupied voxels is, for example, in the literature [J. P It can be determined by applying the method discussed in [Jia-qiang Li et al. in "Target tracking algorithm based on adaptive strong tracking particle filter" (IET Science, Measurement & Technology 10.7, 2016, pp. 704-710)] or in "Factor Graph based 3D Multi-Object Tracking in Point Clouds" (2020 IEEE / RSJ International Conference on Intelligent Robots and Systems (IROS), Las Vegas, NV, USA, 2020, pp. 10343-10350)] or in the literature [Jia-qiang Li et al. in "Target tracking algorithm based on adaptive strong tracking particle filter" (IET Science, Measurement & Technology 10.7, 2016, pp. 704-710)].

[0039] Information regarding the positional change of one or more occupied voxels between the first and second representations of the VEM of the ROI may be expressed in any other suitable form, but in one or more embodiments, a motion prediction vector is determined for one or more voxels representing the positional change between the first and second representations of the VEM of the ROI in the AVI. As mentioned above, determining the motion prediction vector may also consider the occupancy probability of the voxels included in the AVI determined in the comparison step.

[0040] In addition, it is conceivable to store the past position changes of occupied voxels and the associated AVI, and to use this past information to determine the prediction vector. Storing past position changes can be regarded as a tracking history and can enable more accurate predictions, particularly for non-linear trajectories. Such motion predictions can be determined only for voxels or clusters having each occupancy probability that exceeds a first predetermined threshold in at least a second representation of the VEM of the ROI, which can reduce the amount of computation without significantly degrading the usefulness and reliability of the prediction.

[0041] In at least the second representation of the VEM of the ROI, the occupancy probability is lower than a first predetermined threshold but exceeds a second predetermined threshold that is not zero, and the same motion prediction vector as the adjacent voxel can be assigned to the voxel adjacent to the voxel having the occupancy probability exceeding the first predetermined threshold. This can significantly reduce the amount of computation while generating useful prediction results.

[0042] Since uncertainty is expected to be added to the position prediction of voxels in the VEM of the ROI, the occupancy probability of voxels at least before and after the predicted voxel along the corresponding prediction vector can be corrected. Correction can be achieved through addition, subtraction, application of coefficients, averaging, and / or weighting. The amount of correction may vary depending on the occupancy probability of each voxel in one or more second representations of the VEM of the ROI, and a higher occupancy probability may be assigned to predicted voxels that are ahead of the voxels in one or more second representations of the VEM of the ROI in the direction of the prediction vector, and a lower occupancy probability may be assigned to predicted voxels that are behind the voxels in one or more second representations of the VEM of the ROI. The same correction may be applied to voxels adjacent to the predicted voxel to which the motion prediction vector of the predicted voxel has been assigned. In particular, adjusting the occupancy probability of these adjacent voxels to a higher value allows for preparation against inaccuracies in prediction vectors or situations where the orientation of objects represented by voxel clusters changes unpredictably, especially when the voxel clusters are associated with non-stationary objects.

[0043] Identifying a voxel or voxel cluster associated with a moving or stationary object can be improved by providing information on changes in the position of one or more receivers or their antennas that provide one or more first and / or second representations of the VEM of the ROI. This information may be obtained from sensors on a mobile platform equipped with an antenna, e.g., a vehicle or a portable device. Sensors may include speedometers, accelerometers, position information systems, etc. Accordingly, in one or more embodiments of the method, such information is received and provided as additional input to a comparison step and / or a generation step. It is evident that changes in the position of one or more receivers or their antennas must be correlated with the time at which each first and second representation is determined.

[0044] In one or more embodiments, the method further comprises the steps of classifying voxels in the predicted VEM of the ROI based on their respective occupancy probabilities, and merging adjacent voxels belonging to at least one class into corresponding coherent merged voxels. The classification may include at least two classes, in which case a single threshold of occupancy probability is used as the classification criterion. Voxels with an occupancy probability lower than the threshold may be merged, while voxels with an occupancy probability higher than the threshold may be excluded from merging. Generally, the classification may merge voxels with similar occupancy probabilities into the same class. The merging of voxels may include a machine learning-based clustering algorithm. Coherent merged voxels may appear in the predicted VEM of the ROI in an irregular shape or as a rectangular shape.

[0045] A single occupancy probability may be assigned to each coherent merged voxel. This single occupancy probability may be determined by averaging the individual occupancy probabilities of the merged voxels, determining the median of the individual occupancy probabilities of the merged voxels, or determining the root mean square (RMS) value of the individual occupancy probabilities of the merged voxels.

[0046] Voxel merging can be performed for each predicted VEM of an ROI determined from a set of one or more first representations and one or more second representations of the VEM of an ROI, and as a result, changes in the configuration of merged voxels in subsequent predicted VEMs are possible, thereby allowing for the consideration of changes in the location of non-stationary objects within the ROI.

[0047] Voxel merging can reduce the computational load required when later using the ROI's VEM and also decrease storage space requirements. For example, if a merged voxel is marked as unoccupied, it can be considered fully open to radio waves, and channel estimation using the ROI's VEM as input can apply coarser metrics to signals known to pass through this merged voxel. Similarly, merged voxels marked as unoccupied or with a very low probability of occupancy imply a strong assumption of ample open space, which can be utilized for vehicle path generation or path planning.

[0048] As previously mentioned above, the predicted VEM of an ROI can be used in various applications, including a device that executes a process for detecting symbols in wireless communication, a device that executes an environment recognition process, and a device configured to provide an environment representation based on sensor fusion. Accordingly, according to a second aspect of the present invention, a device is provided comprising at least one first interface configured to receive or retrieve one or more first and second representations of the VEM of an ROI, or configured to receive a wireless signal transmitted by one or more transmitters to or across the ROI and determine one or more first and second representations of the VEM of the ROI therefrom, and further comprising at least one second interface configured to output the predicted VEM of the ROI, one or more microprocessors, and associated volatile and non-volatile memory. The at least one first and second interface may share at least partially the same hardware components or may include distinct hardware components. In the former case, the interfaces may be logically separated. However, it is also possible to conceive that the first and second interfaces of the device are identical and operate bidirectionally. The aforementioned components of the device are coupled to be communicable through one or more data lines or data buses. A non-volatile memory includes computer program instructions configured to execute one or more embodiments of the method according to the first aspect of the present invention when executed by at least one of one or more microprocessors.

[0049] Where at least one second interface of the device is configured to receive or retrieve one or more first and second representations of the VEM of the ROI determined from one or more first and second sets of transmitted symbols received at one or more corresponding first and second times by one or more transmitters to or across the region of interest and at one or more receivers, the interface may simply be a digital communication interface, and the data is received according to any suitable digital communication protocol, whether standardized or proprietary. In this case, the determination of the VEM of the ROI has already been performed by another process and possibly another device coupled to the device currently described in a communicable manner. Receiving or retrieving one or more first representations may involve accessing a storage device that stores such information. In this case, the predicted VEM of the ROI provided by executing an embodiment of the method according to the first aspect of the invention presented above may be provided to another device to be used as a preliminary input in determining one or more VEMs of the ROI, provided that each process allows or requires such preliminary input.

[0050] Where at least one first interface of the device is configured to receive a radio signal transmitted by one or more transmitters to or across the ROI, the non-volatile memory includes computer program instructions that further configure the device to determine corresponding first and second VEMs of the ROI before executing one or more embodiments of the method according to the first aspect of the invention presented above, when executed by at least one of one or more processors. Such determination may include, for example, the method described in German Patent Publication No. DE 10 2022 212 615 A1 (the whole of which is incorporated herein by reference). In this case, the predicted VEM of the ROI provided by executing an embodiment of the method according to the first aspect of the invention presented above may be used as a first representation of the VEM of the ROI in a subsequent execution of the method.

[0051] According to a third aspect of the present invention, a first method for operating a wireless communication device is provided. The device includes a third interface configured to receive a predicted VEM of an ROI by executing an embodiment of the method according to the first aspect of the present invention, and a fourth interface configured to receive a wireless communication signal. The method includes the step of receiving a predicted VEM of an ROI at the third interface at a present or future point in time when the reception of a wireless communication signal is scheduled or expected. The predicted VEM of an ROI may be received from a device according to a second aspect of the present invention. Alternatively, the predicted VEM of an ROI may be determined by an additional process executed in the device, for example, a process implementing one or more embodiments of the method according to the first aspect of the present invention. The method further includes the step of determining one or more signal paths in which a wireless communication signal is received or will be received based on the received VEM of an ROI, and the step of predicting channel coefficients for the determined signal paths. Channel count prediction may include iterative or non-iterative methods comprising the process disclosed in patent publication number DE 10 2023 202 288 A1 (the entire contents of which are incorporated herein by reference).

[0052] The method further comprises the steps of receiving a wireless communication signal at a fourth interface, and providing a predicted channel coefficient for each determined signal path as an additional input to a process for estimating a channel coefficient from the received wireless communication signal. Alternatively or additionally, the predicted channel coefficient for each determined signal path may be provided to a process for determining and / or reconstructing a transmission symbol from the received wireless communication signal. Finally, the estimated channel coefficient and / or the determined and / or reconstructed transmission symbol are provided as an output.

[0053] In one or more embodiments of the method according to the third aspect of the present invention, the method further comprises the step of determining a velocity vector of a transmitter receiving a signal for an antenna receiving a signal, based on the received VEM of the ROI and / or the associated AVI. Information represented by the velocity vector may be provided to a process for estimating channel coefficients from a received radio communication signal and / or a process for determining and / or reconstructing transmission symbols from a received radio communication signal. In the latter case, information represented by the velocity vector may be used, for example, to predict and compensate for Doppler shift or spreading of the received signal. Such compensation may include appropriately setting parameters of the receiver's signal detection process according to the expected Doppler shift or spreading.

[0054] According to a fourth aspect of the present invention, a second method for operating a wireless communication device is provided. The device comprises a third interface configured to receive a predicted VEM of a predicted ROI by executing an embodiment of the method according to the first aspect of the present invention, a fifth interface configured to receive a transmission symbol to be transmitted, and a sixth interface configured to transmit a wireless communication signal. The method includes the step of receiving a predicted VEM of an ROI at the third interface at a present or future time when the transmission of a wireless communication signal is scheduled. As with the method according to the third aspect of the present invention, the predicted VEM of an ROI may be received from the device according to the second aspect of the present invention. Alternatively, the predicted VEM of an ROI may be determined by an additional process executed in the device, for example, by a process implementing one or more embodiments of the method according to the first aspect of the present invention and using a previously received wireless signal as input. The method further includes the step of determining one or more signal paths through which a wireless communication signal is transmitted or will be transmitted based on the received VEM of an ROI, and the step of predicting channel coefficients for the determined signal paths. Similar to the method according to the third aspect of the present invention, predicting channel coefficients may include an iterative or non-iterative method comprising the process disclosed in Patent Publication No. DE 10 2023 202 288 A1. The method further comprises the steps of receiving a transmission symbol to be transmitted, and providing the predicted channel coefficients for each determined signal path as additional input to a process of pre-equalizing a transmission signal representing the transmission symbol after coding and / or modulation before transmission. Finally, the pre-equalized transmission signal is transmitted.

[0055] It is evident that the methods according to the third and fourth embodiments can be executed on the same device, for example, when the device is configured for bidirectional wireless communication. In this case, the fourth and fifth interfaces may share hardware components, for example, one or more antennas, radio frequency (RF) amplifiers, modulators, oscillators, etc. Likewise, the third interface may be implemented as a single shared interface.

[0056] The channel estimation used in the method according to the third and fourth embodiments may apply methods including those disclosed in a patent application or patent literature claiming priority of patent publication number DE 10 2023 202 288 A1 and / or German patent application number 10 2023 106 237.9 (the whole contents of which are incorporated herein by reference).

[0057] According to a fifth aspect of the present invention, a wireless communication device is provided. The wireless communication device includes a third interface configured to receive a VEM of an ROI, for example, to receive a predicted VEM of an ROI at a present or future point in time where the reception of a wireless communication signal is scheduled or expected. A fourth interface and / or a fifth interface for receiving a wireless communication signal is likewise provided to the device. The fifth interface is configured to receive a transmission symbol to be transmitted through a sixth interface of the device configured to transmit a wireless communication signal. The wireless communication device further includes one or more microprocessors and associated volatile memory and non-volatile memory. Various components and elements of the wireless communication device are connected via one or more data and / or signal lines or buses. The non-volatile memory includes computer program instructions that configure the wireless communication device to execute the method according to the third aspect of the present invention and / or the method according to the fourth aspect of the present invention when executed by at least one of the one or more microprocessors.

[0058] The third interface of the wireless communication device may provide a communication connection with an external source or may be located within the device. In the latter case, the device may be further configured to perform an embodiment of the method according to the first aspect of the present invention.

[0059] The fourth and fifth interfaces may include and share hardware components, such as one or more antennas, radio frequency (RF) amplifiers, mixers, modulators, oscillators, radio frequency signal processing circuits, etc., and may be configured to receive and / or transmit, respectively, according to a wireless communication protocol or standard.

[0060] The method presented above may be expressed as computer program instructions of a computer program product. Accordingly, according to a third aspect of the present invention, the computer program product includes computer program instructions that, when executed by a processor of a device according to a second aspect of the present invention, cause the device to execute one or more embodiments of the method according to a first aspect of the present invention.

[0061] Computer program instructions may be retrieveably stored or transmitted on a computer-readable medium or data carrier. Such media or data carriers may be physically implemented in the form of, for example, a hard disk, a solid-state disk, a flash memory device, etc. However, such media or data carriers may also include modulated electromagnetic signals, electrical signals, or optical signals that are received by a computer through a corresponding receiver and transmitted and stored in the computer's memory.

[0062] The method presented herein is applicable to any scenario in which an environment is modeled using a voxelized grid model, i.e., a scenario that can be considered in an environment of integrated communication and sensing technology. In particular, the present invention can be used in communication and environmental sensing in indoor and outdoor scenarios using a stationary AP, by a moving vehicle communicating with a Roadside Unit (RSU) while performing vehicle / pedestrian detection, by multiple vehicles cooperating to sense the environment and road conditions without an RSU, or by multiple connected UEs (Bluetooth, Wi-Fi, IoT, etc.) passively sensing the environment (i.e., without using the detection of specific signals). The use cases and benefits of the communication and environmental sensing functions presented herein can each be contained within and achieved within the same wireless interface.

[0063] The present invention improves the modeling accuracy of a VEM by dynamically updating voxelized grid occupancy probabilities, including determined or predicted target movements and other target information, such as available paths and each attribute of these paths, signal scattering in the voxels of the predicted VEM, or uncertainties based on the target's velocity; this information can be used for channel equalization and symbol reconstruction. Furthermore, the present invention improves computational efficiency by dynamically adjusting and refining the grid size based on probabilistic updates. The computational complexity required for continuous environment sensing and the amount of memory required to execute the method are significantly reduced by reducing the resolution of regions of low interest—that is, the voxel size—compared to the resolution of other regions.

[0064] In addition, when the refined VEM provided by the present invention is used as a preliminary input for an iterative step to determine the VEM from a subsequently received signal, the enhanced accuracy of the refined VEM, for example, accuracy considering path blocking, scattering, non-ideal geometric behavior, unrealistic reflection angles, energy loss during scattering, frequency-dependent behavior, and channel path loss based on geometric shape, can help increase the convergence speed of this process. Brief explanation of the drawing

[0065] The attached drawings are used to describe various aspects of the present invention in detail. In the drawings, Figure 1 illustrates an exemplary environment including an object within a region of interest and a voxelized representation of the object. Figure 2 illustrates a representation of the general concept of LOS and NLOS paths in a voxelized space. Figure 3 illustrates a schematic representation of 3D space considered in known systems and methods. Figure 4 illustrates a schematic representation of a communication path that is not available and unrealistic in a voxelized space. FIG. 5 illustrates a schematic and simplified representation of the input of a method according to a first embodiment of the present invention and the output according to a basic embodiment thereof. FIG. 6 illustrates a schematic and simplified representation of the input and output according to a more advanced mode of the method discussed with reference to FIG. 5. FIG. 7 illustrates a simplified overview of the proposed method according to a first aspect of the present invention in that context. FIG. 8 illustrates an exemplary flowchart of a method according to a first embodiment of the present invention. FIG. 9 illustrates an exemplary flowchart of a method according to a third aspect of the present invention. FIG. 10 illustrates an exemplary flowchart of a method according to a fourth aspect of the present invention. FIG. 11 illustrates an exemplary and schematic block diagram of an apparatus according to a second aspect of the present invention. FIG. 12 illustrates an exemplary and schematic block diagram of a wireless communication device according to a fifth aspect of the present invention. Specific details for implementing the invention

[0066] Figures 1 to 4 have already been described in detail above, so they will not be described again.

[0067] FIG. 5 illustrates a schematic and simplified representation of the input of the method (100) according to the first embodiment of the present invention and the output according to the basic embodiment thereof. The simplification is to use a two-dimensional ROI instead of a three-dimensional one. However, the concept is the same regardless of the dimension. FIG. 5a) illustrates the VEM of the ROI set prior to the currently estimated VEM of the ROI shown in FIG. 5b). In the terminology of the present method, FIG. 5a) illustrates a first representation of the VEM of the ROI, and FIG. 5b) illustrates a second representation of the VEM of the ROI. In the example, it is assumed that two voxels are occupied in the previous VEM of the ROI and the occupancy probability is 1. However, it should be noted that the occupancy probability may be lower than 1, and adjacent voxels may also be occupied, but the occupancy probability may be lower. The currently estimated VEM of the ROI shows some uncertainty regarding voxel occupancy, where two voxels have a relatively high probability of occupancy, but their adjacent voxels have a low probability of occupancy. This uncertainty may arise, for example, due to the movement of an object represented by an occupied voxel, because the current estimate was taken at a later point in time than the previously established VEM of the ROI and there is no accurate information regarding the motion vector. The prior and current estimates may be obtained by applying any known method, including the method presented in DE 10 2022 212 615 A1, insofar as it enables the generation of a voxelized environment representation. It should be noted that, as long as this VEM can be converted into a common grid-scale representation by, for example, scaling, voxel-binning, or other appropriate methods, it is irrelevant whether the prior and current estimated VEMs of the ROI were generated through the same device and / or method. The updated or predicted VEM of the ROI is determined according to the present method. This updated or predicted VEM of the ROI can be targeted to be valid at a later point in time than when the ROI's current VEM was determined.This can be useful, for example, in cases where the determination of the ROI's VEM is not consistently performed for any reason, making the ROI VEM at the target time point unavailable. Such situations may occur in mobility scenarios where the object's position is likely to have changed, and the radio receiver using the ROI's VEM to adjust parameters for channel estimation and equalization processes relies on the best possible information regarding the ROI's properties. The updated or predicted output of the method is illustrated in Fig. 5c). Here, the object's motion, represented by the occupied voxels, was determined from the ROI's previous and current VEMs, and this was used to predict or update the ROI's VEM at the current or future time point. As further mentioned above, motion can be determined using generally known motion prediction methods, including object tracking and machine learning (ML)-based and artificial intelligence (AI)-based methods. For prediction, the uncertainty of the voxels cannot simply be shifted according to the motion vector; it must be adjusted based on the ROI's current estimated VEM and the motion vector. In this diagram, the occupancy probability of a voxel is displayed as a pattern filling the voxels, and a darker color indicates a higher occupancy probability.

[0068] Figure 6 illustrates a schematic and simplified representation of the input and output according to a more advanced embodiment of the method discussed with reference to Figure 5. The inputs shown in Figures 6a) and 6b) are assumed to be identical to the inputs discussed in Figures 5a) and 5b). The relevant part of the output, namely the voxels having an occupancy probability higher than a predetermined threshold, is identical to that shown in Figure 5c). However, as an additional optimization, adjacent voxels having a sufficiently low occupancy probability are grouped. In this figure, the empty voxels in the top two rows and the voxels in the rightmost four columns of the bottom two rows are grouped into larger voxels. It should be noted that the voxels are grouped into 2x4 voxels at the top of the VEM and into 2x2 voxels at the bottom of the VEM, which are expressed as multiples of the normal voxel size.

[0069] FIG. 7 illustrates a simplified overview of a proposed method (100) according to a first aspect of the present invention in that context. An input signal (I1) receives one or more first representations of the VEM of the ROI, and an input signal (I2) receives one or more second representations of the VEM of the ROI. One or more second representations of the VEM of the ROI received at input (I2) may be determined via a JCAS method, for example, as disclosed in German Patent Publication No. DE 10 2022 212 615 A1. Other methods for outputting the VEM of the ROI may also be used. The signal received at input (I1) may be provided as an additional input to the method or methods providing one or more second representations of the VEM of the ROI, as indicated by the dashed line. The method (100) further receives one or more uncertainty thresholds at input (I4), which may be used for voxel merging. Then, the predicted VEM of the ROI is output at output (O1).

[0070] FIG. 8 illustrates an exemplary flowchart of a method (100) according to a first aspect of the present invention. In step (110), one or more transmitters transmit to or across a region of interest and obtain one or more first representations of a voxelized environment model of a region of interest determined from or based thereon one or more first transmission symbol sets received at one or more corresponding first time points at one or more receivers. In step (120), one or more transmitters transmit to or across a region of interest and obtain one or more second representations of a voxelized environment model of a region of interest determined from or based thereon one or more second transmission symbol sets received at one or more corresponding second time points different from each first time point at one or more receivers. In step (130), one or more first representations of a voxelized environment model of a region of interest and one or more second representations of a voxelized environment model of a region of interest are compared with each first and second time point to extract or derive auxiliary voxel information. In step (140), a predicted voxelized environment model of the region of interest is generated based on one or more second representations of the voxelized environment model of the region of interest and auxiliary voxel information extracted / derived in step (130). In step (150), the predicted voxelized environment model of the region of interest is output to a process for detecting transmitted symbols, a process for determining the environment model of the region of interest based on received transmitted symbols, and / or a sensor fusion process for determining the environment model of the region of interest.

[0071] FIG. 9 illustrates an exemplary flowchart of a method (300) according to a third aspect of the present invention. The method relates to operating a wireless communication device (500) comprising a third interface (510) configured to receive a VEM of an ROI predicted by the method according to a first aspect of the present invention, and further comprising a fourth interface (520) for receiving a wireless communication signal. In step (310) of the method, a predicted VEM of an ROI at a present or future point in time where the reception of a wireless communication signal is scheduled or expected is received. In step (320), one or more signal paths where the wireless communication signal is received or will be received are determined based on the received predicted VEM of the ROI. Then, in step (330), channel coefficients for each determined signal path are predicted using the received and / or determined information. In step (340), a wireless communication signal is received at the fourth interface (520) of the device (500). For each determined signal path, the predicted channel coefficient is provided as an additional input to the step (350) of estimating the channel coefficient from the received wireless communication signal. Additionally, or alternatively, the predicted channel coefficient for each determined signal path is provided to the step (360) of determining and / or reconstructing the transmission symbol from the received wireless communication signal. In step (370), the estimated channel coefficient and / or the determined and / or reconstructed transmission symbol are output.

[0072] FIG. 10 illustrates an exemplary flowchart of a method (400) according to a fourth aspect of the present invention. The method relates to operating a wireless communication device (500) comprising a third interface (510) configured to receive a VEM of an ROI predicted by the method according to a first aspect of the present invention, a fifth interface (530) configured to receive a transmission symbol to be transmitted, and a sixth interface (540) configured to transmit a wireless communication signal. In step (410) of the method, a predicted VEM of an ROI at a present or future point in time where the transmission of a wireless communication signal is scheduled. In step (420), one or more signal paths to which the wireless communication signal is transmitted or will be transmitted are determined based on the received predicted VEM of the ROI. Then, in step (430), channel coefficients for each determined signal path are predicted using the determined information. In step (440), a transmission symbol to be transmitted is received at the fifth interface (540). A transmission signal representing a transmission symbol after coding and / or modulation is pre-equalized in step (450) before transmitting the pre-equalized transmission signal through the sixth interface (540) in step (460).

[0073] FIG. 11 illustrates an exemplary and schematic block diagram of a device (200) according to a second aspect of the present invention. The device (200) includes at least one first interface (210) configured to receive or retrieve one or more first and second representations of a VEM of an ROI, or configured to receive a radio signal transmitted by one or more transmitters to or across an ROI and determine one or more first and second representations of a VEM of an ROI from the same. The device (200) further includes at least one second interface (220) configured to output a predicted VEM of an ROI, one or more microprocessors (230), and associated volatile memory (240) and non-volatile memory (250). Some or all of the aforementioned components and elements are communicably connected via one or more data lines or buses (260). The non-volatile memory (250) includes computer program instructions that configure the device (200) to execute the method (100) according to the first embodiment of the present invention when executed by at least one of one or more microprocessors (230).

[0074] FIG. 12 illustrates an exemplary and schematic block diagram of a wireless communication device (500) according to a fifth aspect of the present invention. The device (500) includes a third interface (510) configured to receive VEMs of an ROI. The device further includes a fourth interface (520) for receiving wireless communication signals and / or a fifth interface (530) configured to receive transmission symbols to be transmitted, and a sixth interface (540) configured to transmit wireless communication signals. The device further includes one or more microprocessors (550) and associated volatile memory (560) and non-volatile memory (570). Some or all of the aforementioned components and elements are communicably connected via one or more data lines or buses (580). The non-volatile memory (570) includes computer program instructions that configure a wireless communication device (500) to execute a method (300) according to a third aspect of the present invention when executed by at least one of one or more microprocessors (550), and / or configure a device (500) to execute a method (400) according to a fourth aspect of the present invention. Explanation of the symbols

[0075] 100: VEM handling method 110: Acquire first representation 120: Acquiring the second representation 130: Comparison of the First and Second Expressions 140: Predicted VEM generation 150: Predicted VEM output 200: Wireless communication device 210: First Interface 220: Second Interface 230: Microprocessor(s) 240: Volatile memory 250: Non-volatile memory 260: Data line / bus 300: Method 310: Received Predicted VEM 320: Determine signal path(s) 330: Channel coefficient prediction 340: Reception of wireless communication signals 350: Estimation of channel coefficients from received wireless communication signals 360: Transmission Symbol Determination / Reconstruction 370: Estimated channel coefficients and / or determined / reconstructed transmit symbol output 400: Method 410: Received Predicted VEM 420: Determine signal path(s) 430: Channel coefficient prediction 440: Receive transmitted symbol 450: Transmission signal pre-equalization 460: Transmit pre-equalized transmission signal 500: Wireless device 510: The Third Interface 520: The 4th Interface 530: The Fifth Interface 540: The 6th Interface 550: Microprocessor(s) 560: Volatile memory 570: Non-volatile memory 580: Data line / bus

Claims

Claim 1 A method (100) for processing a voxelized environment model of a region of interest for use in wireless communication signal processing and / or environment recognition or mapping, comprising: - a step (110) of obtaining one or more first representations of a voxelized environment model of a region of interest (ROI) determined from or based thereon, one or more first transmission symbol sets received at one or more corresponding first time points by one or more transmitters transmitting to or across the region of interest and at one or more receivers; - a step (120) of obtaining one or more second representations of a voxelized environment model of a region of interest determined from or based thereon, one or more second transmission symbol sets received at one or more corresponding second time points different from each of the first time points by one or more transmitters transmitting to or across the region of interest and at one or more receivers; - a step (130) of comparing one or more first representations of a voxelized environment model of a region of interest and one or more second representations of a voxelized environment model of a region of interest, taking into account each of the first and second time points, to extract or derive auxiliary voxel information (AVI); - one or more second A method comprising: a step (140) of generating a predicted voxelized environment model of the region of interest based on the expression and the auxiliary voxel information extracted or derived in the comparison step (130); and a step (150) of outputting the predicted voxelized environment model (VEM) of the region of interest to a process for detecting, for example, a transmission symbol, a process for determining the environment model of the region of interest based on the received transmission symbol, and / or a sensor fusion process for determining the environment model of the region of interest. Claim 2 In claim 1, the extracted or derived AVI includes information on the position change of one or more occupied voxels between the first and second representations of the VEM of the ROI, in a method (100). Claim 3 A method (100) according to claim 1 or 2, further comprising the step of determining a motion prediction vector for one or more voxels, wherein the AVI represents a position change between the first and second representations of the VEM of the ROI. Claim 4 A method (100) further comprising, in one or more of claims 1 to 3, a step of receiving information about a change in the position of one or more receivers or their antennas, wherein one or more first and / or second representations of the VEM of the ROI are provided, and a step of providing the received information as an additional input to the comparison step (130) and / or the generation step (140). Claim 5 A method comprising, in one or more of claims 1 to 4, a step of classifying voxels in the predicted VEM of the ROI based on their respective occupancy probabilities, and a step of merging adjacent voxels belonging to at least one class into a corresponding coherent merged voxel. Claim 6 A device (200) comprising at least one first interface (210) configured to receive or retrieve one or more first and second representations of a VEM of an ROI, or configured to receive a radio signal transmitted by one or more transmitters to or across the ROI and determine therefrom the one or more first and second representations of a VEM of the ROI, wherein the device further comprises at least one second interface (220) configured to output a predicted VEM of the ROI, one or more microprocessors (230), and associated volatile memory (240) and non-volatile memory (250), wherein the non-volatile memory (250) comprises computer program instructions configured to execute the method (100) of one or more of claims 1 to 5 when executed by at least one of the one or more microprocessors (230). Claim 7 A method (300) for operating a wireless communication device (500) comprising a third interface (510) configured to receive a VEM of a predicted ROI according to the method of one or more of claims 1 to 5, and a fourth interface (520) for receiving a wireless communication signal, wherein the method (300) comprises: - receiving a predicted VEM of an ROI at a present or future point in time at the third interface (510) at which reception of a wireless communication signal is scheduled or expected (310); - determining one or more signal paths to which the wireless communication signal is received or will be received based on the received predicted VEM of the ROI (320); - predicting a channel coefficient for each determined signal path (330); - receiving a wireless communication signal at the fourth interface (520) (340); - a process (350) for estimating the channel coefficient predicted for each determined signal path from the received wireless communication signal and / or determining and / or reconstructing a transmission symbol from the received wireless communication signal. A method comprising the step of providing additional input to the process (360), and the step (370) of outputting the estimated channel coefficient and / or the determined and / or reconstructed transmission symbol. Claim 8 A method (300) further comprising, in claim 7, a step of determining a velocity vector of a transmitter receiving a signal for an antenna receiving a signal based on a received VEM and / or an associated AVI of the ROI, and a step of providing information represented by the velocity vector to a process (360) for estimating the channel coefficient from the received wireless communication signal and / or a process for determining and / or reconstructing a transmission symbol from the received wireless communication signal. Claim 9 A method (400) for operating a wireless communication device (500) comprising a third interface (510) configured to receive a VEM of a predicted ROI according to the method of one or more of claims 1 to 5, a fifth interface (530) configured to receive a transmission symbol to be transmitted, and a sixth interface (540) configured to transmit a wireless communication signal, wherein the method (400) comprises: - receiving a predicted VEM of an ROI at a present or future point in time at which the transmission of a wireless communication signal is scheduled at the third interface (510) (410); - determining one or more signal paths to which the wireless communication signal is transmitted or will be transmitted based on the received VEM of the ROI (420); - predicting a channel coefficient for each of the determined signal paths (430); - receiving a transmission symbol to be transmitted at the fifth interface (540) (440); - pre-equalizing the transmission signal representing the transmission symbol after coding and / or modulation before transmission (460) (450); - the pre-equalized transmission A method comprising the step (460) of transmitting a signal through the sixth interface (540). Claim 10 A wireless communication device (500) comprising a third interface (510) configured to receive a VEM of ROI, and further comprising a fourth interface (520) for receiving a wireless communication signal and / or a fifth interface (530) configured to receive a transmission symbol to be transmitted and a sixth interface (540) configured to transmit a wireless communication signal, and further comprising one or more microprocessors (550) and associated volatile memory (560) and non-volatile memory (570), wherein the non-volatile memory (570) comprises computer program instructions configured to execute the method (300) of claim 7 or 8 and / or the method (400) of claim 9 when executed by at least one of the one or more microprocessors (550). Claim 11 A computer program product comprising computer program instructions configured to execute the method of one or more of claims 1 to 5 when executed by the processor (230) of the device (200) according to claim 6, or configured the device (500) to execute the method of claim 7 or 8 and / or execute the method of claim 9 when executed by the processor (550) of the device (500) according to claim 10. Claim 12 A computer-readable medium or data carrier that transmits or stores the computer program product of paragraph 11 in a retrievable manner.

Images