Radar SDK lets automotive radar sensors see better
NXP Semiconductors is introducing a first version of powerful radar signal processing algorithms as part of its new Premium Radar SDK (PRSDK). These enable designers to optimise the performance of radar systems by tightly coupling NXP’s software algorithms with the latest S32R4x radar processor family. This enables them to further improve the safety and differentiation of radar detection. The library of proprietary radar algorithms enables the integration of ready-made and already validated algorithms into radar sensor applications. This helps accelerate radar sensor development and reduce R&D investment. The first version will be ready for evaluation during 2022. It includes three algorithm packages with solutions for interference cancellation, MIMO waveform optimisation and angular resolution improvement.
The background is that radar applications such as automatic emergency braking and adaptive cruise control are increasingly becoming standard in vehicles delivered today. Regional laws and guidelines, as well as the Euro NCAP criteria, prescribe increasingly sophisticated functions for blind spot assist, turn assist and cross-traffic detection of people in front of and behind the vehicle. This requires more radar nodes per vehicle to create a 360-degree safety zone. Also, the growing demand for L2+ autonomy level vehicles, which offers comfort features close to L3, requires more powerful radar sensors to better detect difficult boundary cases and provide accurate environment mapping and localisation.
NXP expects the tripling of growth in the use of automotive radar systems to continue over the next decade. More vehicles will be equipped with radar sensors, the number of sensor nodes per vehicle will increase and more powerful sensors will be deployed. This trend poses several challenges for vehicle OEMs and Tier 1 suppliers. In particular, there is a need for interference mitigation, MIMO waveform optimisation and increased sensor resolution. NXP’s Premium Radar SDK algorithms are designed to address these challenges, explains Huanyu Gu, Director Product Marketing and Business Development ADAS at NXP.
在五年内,他估计,汽车将情商uipped with twice as many radar sensors as today, and over 90% of radar sensors will transmit within the 77-79 GHz band. Concern about radar interference is therefore growing, and the need for remedial action is becoming increasingly urgent. At the same time, car manufacturers are looking for ways to improve sensor performance. Optimised MIMO waveforms are therefore essential to provide radar sensors with higher resolution and longer range. Special modulation processes here allow the simultaneous operation of several transmitters and encode the signals of the individual transmitting antennas. This ensures that the signals can then be distinguished on the receiver side. In addition, sensors with higher resolution are desirable for both corner and front radar applications, as they can be used to achieve more precise object separation and classification for use cases such as the detection of endangered road users or parking assistance.
The Premium Radar SDK is expected to evolve over time, meaning that its algorithms will be regularly adapted to new radar sensor requirements and NXP’s future radar roadmap. The PRSDK implementation complies with international automotive quality management standards IATF 16949:2016 and ASPICE Level 3 requirements, and the products allow for user customisation to meet needs for customisation and differentiation.
NXP offers OEMs and Tier 1 suppliers two options for evaluation licences to evaluate the SDK. A MATLAB-based version provides the algorithms as compiled code to help developers understand how the PRSDK works. To do this, they can feed in test vectors from the client, calculate them and visualise the output vectors. An MCU-based evaluation version, on the other hand, contains the algorithm binaries that can be executed on the evaluation board of the respective NXP Radar processor.
+MORE
