Ensilica Plc · LSE:ENSI

Ensilica — ADAS Sensors 2017 Detroit

8 slides from Ensilica's conference presentation, May 17, 2017

  1. Slide 1: Challenges of Automotive Radar Sensor Networks

    CHALLENGES OF AUTOMOTIVE RADAR SENSOR NETWORKS Dr David Wheeler 17th May, 2017 Page 1

  2. Slide 2: 360° Coverage by Sensor Units

    360° Coverage by Sensor Units your trusted IC partner © 2017 EnSilica Limited | Page 2

  3. Slide 3: 360° Coverage by Sensor Network

    360° Coverage by Sensor Network parking/SRR parking/SRR/LRR 8 ECU Sensor Links ▪ Multi-mode RADAR and parking without ultrasonic sensors. ▪ Central ECU receives and processes RADAR sensor signals ▪ Central ECU performs RADAR and camera sensor fusion your trusted IC partner © 2017 EnSilica Limited |

  4. Slide 4: Challenges of Sensor Network

    Challenges of Sensor Network ▪ Ideally one sensor module design to cover parking, SRR, MRR and LRR ▪ RF modulation supporting a wide distance range ▪ Each sensor module has between 4 and 12 receive antennas ▪ Data rates from each sensor module to ECU ▪ Option 1: 12-channels of 12-bit raw RF samples

  5. Slide 5: Sensor Module data transfer to ECU

    Sensor Module data transfer to ECU Protocol Max bit rate Distance Object list Raw RF FFT CAN 1 Mbit/s 40 m    FlexRay 10 Mbit/s 22 m ✓   BroadR Ethernet 100 Mbit/s 5m ✓✓ 88Q2112 Ethernet 1 Gbit/s 10 m ✓✓✓ MIPI CSI2 + D-PHY 2.5 Gbit/s / lane 30 cm    Proprietary LVDS 4 lane 10 Gbit/s ✓✓✓ ✓ 

  6. Slide 6: Minimum sensor processing every 20 ms

    Minimum sensor processing every 20 ms Processing fixed point MOp/s floating point MOp/s 1024 point Range FFT 6300 256 point Doppler FFT 630 1024 point Az'th and Elevation FFT 525 Range CFAR Detection 100 Doppler CFAR Detection 3350 Kinematics 0.7 Measurement validation 530 Track association 625 Kalm

  7. Slide 7: ADAS RADAR Co-Processing Solution

    ADAS RADAR Co-Processing Solution ▪ Free up the ECU for FUSION ▪ Highest levels of RADAR processing integration ▪ Supports 12 antennas @ 100 Msps with plot extraction and tracking 128 objects ▪ Low power: < 200 mW ▪ Small package: 8 mm x 8 mm your trusted IC partner © 2017 EnSilica Limited | Page 7

  8. Slide 8: Conclusion

    Conclusion Dedicated RADAR co-processing engine enabling: ▪ Parking, short, long range & 3D RADAR ▪ RADAR sensor networks providing object lists to ECU ▪ Lowest latency and power solution ▪ ECU headroom for track-to-object recognition and make safety decisions your trusted IC partner © 2017 EnSilica

EnSilica — ADAS Sensors 2017 Detroit on 6ix