WEpods: Autonomous Shuttles on Public Roads WEpods partners WEpod - - PowerPoint PPT Presentation

▶
wepods autonomous shuttles on public roads wepods
SMART_READER_LITE
LIVE PREVIEW

WEpods: Autonomous Shuttles on Public Roads WEpods partners WEpod - - PowerPoint PPT Presentation

Mar 11, 2023 127 likes •347 views

WEpods: Autonomous Shuttles on Public Roads WEpods partners WEpod route WEpods functional architecture EasyMile EZ10 WEpods global localization WEpods functional architecture IBEO localization ADASIS e-Horizon Extended ADASIS v2 e-Horizon

slide-1
SLIDE 1

WEpods: Autonomous Shuttles on Public Roads

slide-2
SLIDE 2

WEpods partners

slide-3
SLIDE 3

WEpod route

slide-4
SLIDE 4

WEpods functional architecture

EasyMile EZ10

slide-5
SLIDE 5

WEpods global localization

slide-6
SLIDE 6

WEpods functional architecture

Extended ADASIS v2 e-Horizon with lateral position IBEO localization ADASIS e-Horizon localization Road users &

  • bstacles
slide-7
SLIDE 7

WEpods functional architecture

slide-8
SLIDE 8

Sensing - Camera

slide-9
SLIDE 9

Sensing - Camera

slide-10
SLIDE 10

Sensing - Camera

slide-11
SLIDE 11

Sensing - Field of View

slide-12
SLIDE 12

Sensing – Radar-Camera combination

  • Radar detection

× Unknown type of object  Location of object  Low false negative rate

  • Visual (pedestrian) detection

× Processing of whole image × Unknown visual scale × High false positive rate × Weather conditions etc.  Recognition of object

slide-13
SLIDE 13

Sensing – Radar-Camera combination

  • Radar-Camera Detection & Classification

 Location of object  Projection to camera view  Recognition of object

  • System Architecture
  • Setup on DrivePX
  • Radar to Camera projection and visual cropping
  • Deep-learned Convolutional Neural Network
  • Network architecture
  • Network training
  • Tracking and fusion
slide-14
SLIDE 14

Sensing – Setup

  • Setup on DrivePX:
  • Radar inputs over Aurix CAN interface
  • Camera inputs over CSI interface
  • Cropping based on radar to camera projection
  • NVidia CUDA and CuDNN
  • Caffe for DNN-library
slide-15
SLIDE 15

Sensing – Radar-Camera combination

  • Radar – Camera projection
  • Point projection
  • Object distance
  • Camera Calibration
  • Object size
  • Visual scale
slide-16
SLIDE 16

Sensing – Classification

Network architecture:

Conv

9x9 128 filters

Conv

7x7 512 filters

Max

2x2

Conv

3x6 1024 filters

Max

2x2

Conv

3x8 128 filters Image crop 40x100 56x116 Feature learning Class learning Class

slide-17
SLIDE 17

Sensing – Classification

Network training:

  • Robustness to small variations
  • Translation
  • Scale
  • Flip
  • Contrastive loss learning
  • Robustness to appearance variations
  • Contrastive loss learning
  • Boosting
  • Continuous learning
  • Tracking feedback
  • False classifications retrained
slide-18
SLIDE 18

Sensing – Tracking

Tracking:

  • Continuous localization
  • Fusing sources
  • Increasing robustness
  • Short term prediction
slide-19
SLIDE 19

Conclussion

  • Combining radar and camera
  • Deep-learned Convolutional Neural Network
  • Less false positives
  • 3D localization
  • Multiple types of road users
  • Future work
  • Combining Visual lane detection and

localization with e-Horizon

  • Pedestrian intent recognition
  • Road user intent recognition
slide-20
SLIDE 20