Co-processing Engine for Advanced Driver Assistance Systems
The EnSilica eSi-ADAS IP core provides a small, highly efficient solution to enhance overall radar based ADAS system performance and capabilities for automotive, drone and UAV applications requiring fast and responsive situational awareness using multiple sensors.
The compact, low gate-count architecture of the eSI-ADAS core enables the high bandwidth computationally intensive operations involved in plot extraction and tracking to be quickly and efficiently processed at the radar receiver stage, significantly reducing the load and overhead on the main ADAS system CPU /DSP and at the same time boosting overall ADAS capabilities. Operating on range, velocity and angular measurements eSi-ADAS applies advanced digital processing techniques including 3 dimensional Fast Fourier Transforms, burst averaging to improve signal-to-noise ratio, Constant False Alarm (CFAR) detection and Kalman Filtering. These operations all take place in real-time to constantly update up to 128 extracted plots and their associated movement.
eSi-ADAS is a co-processing engine that combines the radar measurements with a dynamic motion model to predict and update target position and velocity estimates providing a generic algorithm framework for the central ADAS processor to fuse measurements from different sensors into a single target track.
To provide the most precise plot extraction eSi-ADAS operates in conjunction with a fast chirp capable RF stage and multiple receive channels. EnSilica has implemented its patented techniques to overcome the disadvantages usually associated with processing fast chirp modulation, notably its high computational and memory requirements.
Compared to most current ADAS processing methods, eSi-ADAS offers:
- Up to 10x lower area & power
- Up to 20x lower memory requirement
- Real-time tracking of up to 128 plots
- Fusion of short range and long range radar modes
- Programmable, efficient hardware solution