Systems and methods for remotely tracking life signs with a millimeter-wave radar
Systems and methods for monitoring life signs in a subject a radar unit comprising generates raw data and a processor unit receives raw data and identifies oscillating signals in a series of frames of complex values representing radiation reflected from each voxel of a target region during a given time segment. A processor may collate a series of complex values for each voxel representing reflected radiation for the associated voxel in multiple frames and generate therefrom a real signal sample for each voxel. Neural networks may be used to select the voxel with the best waveform.
1 . A method for monitoring life signs in a subject within a target region, the method comprising:
providing a radar unit comprising:
at least one transmitter antenna connected to an oscillator, and
at least one receiver antenna operable to generate raw data;
providing a processor unit configured to receive raw data from the radar unit and operable to identify oscillating signals therein;
the radar unit transmitting electromagnetic waves into the target region;
the radar unit receiving electromagnetic waves reflected by objects within the target region; the radar generating a series of frames, each frame comprising an array of complex values representing radiation reflected from each voxel of the target region during a given time segment;
for each voxel, the processor collating a series of complex values representing reflected radiation for the associated voxel in multiple frames;
generating a series of overlapping complex signal segments;
converting each complex signal segment into a feature segment;
transferring feature segments to a contiguity manager;
generating a contiguous real value signal for each voxel;
generating a real signal sample for each voxel;
transferring the real signal samples generated for each voxel to a scoring network;
generating an evaluation score for the real signal samples;
selecting a voxel with highest evaluation score; and
extracting life signs parameters from selected real signal.
2 . The method of claim 1 wherein the step of generating the evaluation score for the real signal samples comprises calculating a similarity score fore each voxel.
3 . The method of claim 1 wherein the step of generating the evaluation score for the real signal samples comprises evaluating a relative score for each voxel.
4 . The method of claim 1 wherein the step of generating the evaluation score for the real signal samples comprises: extracting a set of features from each real signal; and evaluating voxel feature scores.
5 . The method of claim 1 wherein the step of generating the evaluation score for the real signal samples comprises:
calculating a similarity score fore each voxel;
evaluating a relative score for each voxel;
extracting a set of features from each real signal;
evaluating voxel feature scores; and
combining multiple scores.
6 . The method of claim 1 wherein the step of generating a series of overlapping complex signal segments comprises collecting continuous sequences of complex values for each voxel, each continuous sequences being collected from a continuous sequence of frames over a sequence-duration Δt.
7 . The method of claim 6 wherein a stagger-time δt elapses between sampling a first complex value for each complex signal segment and sampling the first complex value of a subsequent complex signal segments.
8 . The method of claim 7 wherein the stagger-time δt is less than the sequence-duration Δt.