Vehicle Pose Estimation using UWB Radios Alireza Ansaripour - - PowerPoint PPT Presentation

β–Ά
vehicle pose estimation using uwb
SMART_READER_LITE
LIVE PREVIEW

Vehicle Pose Estimation using UWB Radios Alireza Ansaripour - - PowerPoint PPT Presentation

Aug 22, 2023 240 likes β€’716 views

ViPER Vehicle Pose Estimation using UWB Radios Alireza Ansaripour (University of Houston) Milad Heydariaan (University of Houston) Omprakash Gnawali (University of Houston) Kyungki Kim (University of Nebraska-Lincoln) DCOSS 2020 June 2020 1

slide-1
SLIDE 1

ViPER Vehicle Pose Estimation using UWB Radios

Alireza Ansaripour (University of Houston) Milad Heydariaan (University of Houston) Omprakash Gnawali (University of Houston) Kyungki Kim (University of Nebraska-Lincoln) DCOSS 2020 June 2020

1

slide-2
SLIDE 2
  • Introduction
  • Challenge
  • Related work
  • Design
  • Evaluation
  • Conclusion

2

ViPER Vehicle Pose Estimation using UWB Radios

slide-3
SLIDE 3

Road construction safety

  • About 773 per year lose their lives in

work zone crashes1.

  • 1982 - 2017
  • Some of these can be prevented
  • Monitoring the location
  • Pose estimating systems
  • Track the entities

3

1 https://www.cdc.gov/niosh/topics/highwayworkzones/default.html

slide-4
SLIDE 4

Requirements for location system for construction safety

  • Consistent location availability
  • All time
  • Real-time
  • Notify the workers quickly as possible
  • Low location estimation error
  • Avoid false alarms

4

slide-5
SLIDE 5

Pose estimation technologies

Non-UWB technologies Multi-sensors technologies UWB-based technologies

  • GPS + IMU
  • Weinstein (2010)
  • Multiple-cameras
  • Soltani (2017)
  • IMU + UWB
  • Strohmeier (2018)
  • GPS + UWB
  • GonzΓ‘lez (2007)
  • Multiple-UWB
  • Zhang (2012)
  • Low accuracy
  • High implementation cost
  • Data fusion problem
  • Non-Line of sight problem

5

slide-6
SLIDE 6

UWB based pose estimation

6

Anchor Tag Data is collected with anchors and tags

slide-7
SLIDE 7

Pose Estimating Systems

7

Output: Location

Workers Vehicles

slide-8
SLIDE 8
  • Introduction
  • Challenge
  • Related work
  • Design
  • Evaluation
  • Conclusion

8

ViPER Vehicle Pose Estimation using UWB Radios

slide-9
SLIDE 9

Boundary estimation (ideal case)

9

Tag placement on vehicle Calculated location of tags

slide-10
SLIDE 10

Boundary estimation (ideal case)

10

Tag placement on vehicle Calculated location of tags

slide-11
SLIDE 11

Boundary estimation (real-world scenario)

11

Calculated Locations

Tag placement on vehicle

slide-12
SLIDE 12

Boundary estimation (real-world scenario)

12

Calculated Locations

Tag placement on vehicle Estimation 1

slide-13
SLIDE 13

Boundary estimation (real-world scenario)

13

Calculated Locations

Tag placement on vehicle Estimation 2

slide-14
SLIDE 14

Boundary estimation (real-world scenario)

14

Calculated Locations

Tag placement on vehicle Estimation 3

slide-15
SLIDE 15

Boundary estimation

  • Inaccurate localization
  • Non-line of sight (NLoS) condition
  • Different possibilities for boundary
  • When using mapping method (trivial method for mapping)

15

slide-16
SLIDE 16
  • Introduction
  • Challenge
  • Related work
  • Design
  • Evaluation
  • Conclusion

16

ViPER Vehicle Pose Estimation using UWB Radios

slide-17
SLIDE 17

Related work

  • Averaging methods (Zhang C. (2012))
  • Reduce the error in estimating the pose
  • Not effective in construction site environments
  • Optimization method (Vahdatikhaki F. (2015))
  • Specific type of vehicles
  • Limited type of movements
  • Data fusion (Strohmeier M. (2018))
  • Sophisticated
  • Limited to simple environments

17

slide-18
SLIDE 18
  • Introduction
  • Challenges
  • Related work
  • Design
  • Evaluation
  • Conclusion

18

ViPER Vehicle Pose Estimation using UWB Radios

slide-19
SLIDE 19

Design overview

  • Localization engine
  • TDoA localization
  • Low-pass filter
  • Anchor and reference selection
  • Pose estimator
  • Removing inaccurate estimates
  • Rectangle optimizer

19

slide-20
SLIDE 20

TDoA Localization

  • Used in our localization engine
  • 1. Collects the received timestamp of the

signal from anchors

  • 2. If more than 4 anchors reported

timestamp for a tag

1. One anchor is chosen to be reference 2. TDoA inputs are calculated 𝐽𝑗 = 𝑑 βˆ— (𝑒𝑗 βˆ’ 𝑒𝑠𝑓𝑔) 3. Calculates the location of the tag 𝐺 𝐽1, … , π½π‘œ β†’ π‘Œ, 𝑍, π‘ π‘“π‘‘π‘—π‘’π‘£π‘π‘š_𝑓𝑠𝑠𝑝𝑠

20

𝐡0 π‘ˆ0 𝐡1 𝑒1 βˆ’ 𝑒0 = 𝐽1 𝐡2 𝑒2 βˆ’ 𝑒0 = 𝐽2 𝑒3 βˆ’ 𝑒0 = 𝐽3 𝐡3 Reference anchor

slide-21
SLIDE 21

Correcting TDoA input

  • 1. Low-pass filter
  • 2. Anchor selection
  • 3. Reference selection

21

slide-22
SLIDE 22

TDoA input observation

  • TDoA input for static tag
  • Expected result to be static
  • 𝐽1 = 𝑒1 βˆ’ 𝑒0
  • Plenty of fluctuations in observation

22

Ground truth 𝐽1(𝑛) Anchor #1 Anchor #0 (reference anchor) Tag 𝑒1 𝑒0

slide-23
SLIDE 23

Correction 1: Low-pass filter

  • Remove the noise in TDoA input
  • Designed a low-pass filter
  • Parameters
  • Cut-off frequency : 5 Hz
  • Order : 5

23

slide-24
SLIDE 24

Correction 1: Low-pass filter

24

Ground truth Ground truth Raw Input Low-pass filtered

Some of the noises were removed by applying low-pass filter on TDoA input Low-pass filter

𝐽1(𝑛) 𝐽1(𝑛)

slide-25
SLIDE 25

Results for a moving tag

25

Low-pass filter

Ground truth

Low-pass filter was able to reduce the number of missing points

slide-26
SLIDE 26

Correction 2: Anchor selection

  • Feed the optimizer with more

validated data

  • Removing inaccurate

measurements

  • More than 4 anchors reporting
  • Gap between actual and filtered

value

  • Gap threshold: 2 meters

26

Low-performance correction Ground truth Filtered result

slide-27
SLIDE 27

Anchor Selection (real-world results)

27

Distance difference (m)

𝐽1

slide-28
SLIDE 28

Correction 3: Reference selection

  • Large number of anchors were

removed by Anchor Selection

  • Error in the time stamp of the

reference

  • Propagate to all TDoA inputs
  • 𝐽𝑗 = 𝑑 βˆ— (𝑒𝑗 βˆ’ 𝑒𝑠𝑓𝑔)
  • Modify the reference anchor
  • One with least number of removed

anchors

28

All anchors are removed by anchor selection Distance difference (m) Sample number 𝐽3 𝐽1 𝐽2

slide-29
SLIDE 29

Reference Selection (real-world results)

29

Reference selection

The number of removed anchors decreased as we changed the reference anchor

Distance difference (m) Distance difference (m)

slide-30
SLIDE 30

Results for a moving tag

30

Anchor and reference selection

Ground truth

slide-31
SLIDE 31

Pose estimation

  • Removing erroneous location
  • Rectangle optimization method

31

slide-32
SLIDE 32

Correction 4: Removing Erroneous Locations

  • Estimate the error of the location
  • Value of the TDoA optimization
  • Residual value
  • Remove the locations
  • High residual value
  • Threshold = 5

32

slide-33
SLIDE 33

Correction 5: Rectangle Optimizer

33

Calculated Locations Tag placement for vehicle

slide-34
SLIDE 34

Correction 5: Rectangle Optimizer

34

(𝑦, 𝑧, πœ„) Initial guess

slide-35
SLIDE 35

Correction 5: Rectangle Optimizer

35

𝑒3,1 (𝑦, 𝑧, πœ„)

slide-36
SLIDE 36

Rectangle Optimizer

  • Objective function

𝑔 𝑦, 𝑧, πœ„ = ෍

𝑗=1 𝑂

෍

π‘˜=1 𝑑𝑗𝑨𝑓(𝑗)

𝑒𝑗,π‘˜

2

  • Finds the location and orientation of the vehicle by

ΰ·  π‘ˆ = π‘π‘ π‘•π‘›π‘—π‘œ

𝑦,𝑧,πœ„

𝑔(𝑦, 𝑧, πœ„)

36

slide-37
SLIDE 37

Correction 5: Rectangle Optimizer

37

Objective value = 1000

slide-38
SLIDE 38

Correction 5: Rectangle Optimizer

38

Objective value = 10

slide-39
SLIDE 39

Correction 5: Rectangle Optimizer

39

Objective value = 1

slide-40
SLIDE 40

Correction 5: Rectangle Optimizer

40

Objective value = 0.05 Residual value = 0.05

slide-41
SLIDE 41
  • Introduction
  • Challenges
  • Related work
  • Design
  • Evaluation
  • Conclusion

41

ViPER Vehicle Pose Estimation using UWB Radios

slide-42
SLIDE 42

Evaluation setup and metrics

  • Environment setup
  • Campus parking lot
  • Line of sight environment
  • No object blocking the signal
  • Road construction site
  • Objects causing NLoS conditions
  • Evaluation metrics
  • Location availability
  • Error rate

42

Campus parking Construction site Anchor placement

slide-43
SLIDE 43

Results (location availability)

43

SOA ViPER Construction site 46% 100% Parking lot 94% 98%

Anchor and reference selection increased the location availability by 117%

slide-44
SLIDE 44

Results (error rate)

44

Rectangle optimization was able to reduce the error rate by 90%

Accepted threshold

Location difference (m) Orientation difference (m)

Differences higher than the accepted threshold is considered as error in our application CDF of difference in location and orientation estimate compared to the ground truth

ViPER SoA ViPER SoA

slide-45
SLIDE 45

Limitations

  • Number of tags
  • Time division medium access approach
  • Based on the update rate of the tag
  • 𝑂𝑣𝑛 𝑝𝑔 π‘‘π‘šπ‘π‘’π‘‘ = π‘‰π‘žπ‘’π‘π‘’π‘“ 𝑠𝑏𝑒𝑓 βˆ— 𝑂𝑣𝑛 𝑝𝑔 𝑒𝑏𝑕𝑑
  • Currently supports 40 tags with update rate of 4
  • Robustness
  • Decrease in accuracy
  • One or more anchors stop sending signal for a long time
  • Average 2-10% drop in accuracy for each tag removed

45

slide-46
SLIDE 46
  • Introduction
  • Challenges
  • Design
  • Evaluation
  • Conclusion

46

ViPER Vehicle Pose Estimation using UWB Radios

slide-47
SLIDE 47

Conclusions

  • Pose estimation system
  • Monitor the safety in construction site environment more accurately
  • Improvements
  • Location reception ratio and error rate
  • Methods
  • Correcting or removing inaccuracy in TDoA inputs
  • Rectangle optimization to enhance boundary estimation

Contact: aansarip@cougarnet.uh.edu

47