Executive Summary : | The development of ultra-high-resolution radars is crucial for reaching level-5 autonomy in autonomous vehicles. Current RF-CMOS-based automotive FMCW radars fail to deliver the desired level of separability of close-by targets in the range-Doppler-angle domain required for level-2+ autonomy. The proposal aims to design an ultra-high resolution FMCW radar system with target separation performance at the required level for achieving level-5 autonomy. The key idea is to collate sub-sampled radar measurements from multiple cascaded radar sensors and process these measurements in a centralized fashion using a digital signal processor. The compressed-sensing powered FMCW radar solution will have several advantages, including being significantly cheaper than LiDAR-based alternatives, having a scalable architecture to support higher range, Doppler, and angular resolution on demand, better at filtering out false alarms, less susceptible to side-lobe clutter, and working seamlessly with existing FMCW radar transceiver architecture. This cost-effective solution will overcome technological and cost-related barriers in achieving level-5 autonomy, accelerating the large-scale deployment of low-cost FMCW radar-equipped fully autonomous vehicles and saving lakhs of lives lost annually in India and worldwide by preventing road accidents. |