Decentralized Radar Sensor Local Processing Reduces Heat

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Radar systems face challenges in reducing processing intensity and heat generation at sensors, leading to increased costs and size, while also requiring high-speed data transfer for multiple sensors, which is expensive and space-constrained.

Innovation Solution

Implementing a centralized host processor to perform more intensive processing, with sensors performing local processing to transform radar signals into distance/relative velocity matrices and filtering samples with high signal-to-noise ratios for transfer, reducing the amount of data sent to the host processor.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Area of stationary object

If multiple radar sensors are deployed to increase object detection coverage, then the coverage area and detection accuracy are improved, but the processing power requirement and system cost increase

Engineering Contradiction:
Improvecoverage areaVSAvoidprocessing power requirement
Core Design Contradiction:
Area of stationary objectVSPower

Solution Approach 1:

The patent segments the processing tasks by separating local processing at each sensor node from centralized processing at the host processor. Each sensor node performs local processing to generate measurement matrices and identifies samples exceeding a threshold, while the host processor performs centralized processing on the reduced data set. This segmentation reduces the processing power requirement at each individual node while maintaining the benefits of multiple sensors for improved coverage area.

Inventive Principle:
Principle #1Segmentation

2Measurement precision

If multiple radar sensors are deployed to increase object detection coverage, then the detection accuracy is improved, but the system cost increases

Engineering Contradiction:
Improvedetection accuracyVSAvoidsystem cost
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent extracts and removes unnecessary data processing steps from the system by performing threshold-based filtering at the sensor node level before data transmission. Only samples exceeding the threshold are transmitted to the host processor, extracting the essential information while discarding redundant data. This reduces the overall system cost by minimizing the processing requirements and data transfer needs while maintaining detection accuracy through the use of multiple sensors.

Inventive Principle:
Principle #2Taking out (Extraction)

3Volume of moving object

If sensors are miniaturized to fit unobtrusively in the vehicle, then the space utilization is improved, but the heat generation problem worsens

Engineering Contradiction:
Improvesensor sizeVSAvoidheat generation
Core Design Contradiction:
Volume of moving objectVSTemperature

Solution Approach 1:

The patent applies partial processing at the sensor node level by performing local processing to generate measurement matrices and applying threshold filtering before data transmission. This partial action reduces the processing load and heat generation at the miniaturized sensor nodes, while the remaining processing is performed centrally at the host processor which has better thermal management capabilities. This allows the system to use miniaturized sensors for improved space utilization while managing heat generation through distributed processing.

Inventive Principle:
Principle #16Partial or excessive action

4Speed

If high-speed data transfer interface is provided for multiple sensors, then the data processing speed is improved, but the cost and space requirements increase

Engineering Contradiction:
Improvedata processing speedVSAvoidinterface cost
Core Design Contradiction:
SpeedVSDevice complexity

Solution Approach 1:

The patent performs preliminary action by conducting threshold-based filtering and sample identification at the sensor node level before data transmission. This preliminary processing reduces the volume of data that needs to be transmitted over the interface, thereby reducing the required data transfer speed and interface complexity. The host processor receives only the essential samples exceeding the threshold, which maintains data processing speed while reducing interface cost and space requirements.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP3324205B1Decentralised radar system
Publication Date: 2023.01.11 NXP BV
  • EP3324205B1 patent drawingFigure 1
  • EP3324205B1 patent drawingFigure 2
  • EP3324205B1 patent drawingFigure 3

AI summary

There is provided a radar sensor and method. The radar sensor comprises a plurality of transmit and receive antennas, a transceiver, a digital signal processor, a filter and an interface. The transceiver is configured to digitize received radar signals to provide a plurality of digital samples. The digital signal processor is configured to form a measurement matrix by transforming the plurality of digital samples into a distance/relative velocity matrix for each combination of the transmit and receive antennas. The filter is configured to identify samples forming the measurement matrix having a signal to noise ratio higher than a threshold value. The interface is configured to transmit the identified samples and their location in the measurement matrix to a remote host processor configured to further carry out direction of arrival processing on the identified samples.