Radar detection software improves accuracy using less computing

Radar detection software improves accuracy using less compute power

Teraki’s ML-detection software on Infineon Technologies’ microcontrollers delivers more points per object, leading to less false positives and increased safety compared to other radar processing techniques, such as CFAR.

Advertisement

Edge sensor processing firm Teraki has released its latest radar detection software it said accurately identifies static and moving objects with increased accuracy and less computational power. The real traffic solution runs on ASIL-D compliant AURIX TC4x microcontrollers from Infineon Technologies.

Autonomous driving (AD) and advanced driver assistance systems (ADAS) rely on precise sensing of a vehicle’s surrounding environment to safely navigate. The role of advanced sensors and algorithms is to enhance perception and ensure safety. Radar is becoming a key technology for cost effective signal processing but there are limitations which need to be overcome.

For example, interference can severely detriment radar detection performance, leading to erroneous detections in difficult multi-target situations, which also carries high processing requirements. In addition, the precision required for reliable radar classifications involves more data points per frame and sub-1-degree angular resolution, if static and moving objects are to be correctly detected and classified.

Teraki said its machine learning (ML) approach intends to solve this challenge by working with raw data and playing both a denoising and a cognitive role in dissecting information from the radar, identifying targets amid noisy environments, clusters, and other interference while decreasing the processing capacity at the edge. The processing pipeline employs ML to reduce the data required to achieve accurate detections while improving the quality and density of data points of individual detections. This ML-detection delivers more points per object, leading to less false positives, and thus to increased safety, compared to other radar processing techniques, such as CFAR (continuous false alarm rate).

Teraki Infineon block diagram
In Teraki’s hybrid approach, the company said it overcomes the challenges of radar processing by combining traditional signal processing with machine learning. (Image: Infineon Technologies)

Ported with Infineon’s AURIX TC4, Teraki’s ML-based algorithm claims to reduce radar signals after the 1st FFT achieving up to 25x lower error rates of missing objects at the same RAM/fps. Compared to CFAR, classification is up to 20 percent higher in precision, and valid detections increase 15 percent more. With this release, Teraki said it is improving the chipset architecture of edge devices, ensuring real-time processing performance on AURIX TC4, alleviating the computing requirements by consuming 4- or 5-bit bitrates instead of 8- or 32bits without compromising the F1-scores. This leads to up to 2 times less memory required.

Teraki CEO Daniel Richart said, “We have refined our software to achieve more with less. Our solution allows to better detect and to correctly classify static and moving objects from radar signals. In addition, it enables customers to detect obstacles at farther distances. This provides AD- and ADAS-applications with more reliable information and hence better situational awareness that leads to safer decision-making. Ultimately, we ensure safety by reducing inference time and the required processing power at the edge.”

The director of product marketing for Infineon automotive microcontrollers, Marco Cassol, said that automotive radar system performance has drastically increased over the last product generation. He added, “Edge AI processing is one of the many innovations that has helped us drive this increase in radar performance. Infineon’s new parallel processing unit (PPU) are now being implemented in Teraki’s unique radar algorithms to showcase next-generation radar performance from Infineon’s AURIX TC4x devices.”


Related Content:

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.