An Edge-Cloud System Model for Autonomous Vehicles Yu Sasaki - - PowerPoint PPT Presentation

▶
an edge cloud system model for autonomous vehicles
SMART_READER_LITE
LIVE PREVIEW

An Edge-Cloud System Model for Autonomous Vehicles Yu Sasaki - - PowerPoint PPT Presentation

Sep 12, 2023 304 likes •549 views

An Edge-Cloud System Model for Autonomous Vehicles Yu Sasaki ,Tomoya Sato , Hiroyuki Chishiro ,Tasuku Ishigooka , Satoshi Otsuka , Kentaro Yoshimura , Shinpei Kato Dept. Computer Science, The University of Tokyo,

slide-1
SLIDE 1

An Edge-Cloud System Model for Autonomous Vehicles

Yu Sasaki† ,Tomoya Sato†, Hiroyuki Chishiro† ,Tasuku Ishigooka§, Satoshi Otsuka §, Kentaro Yoshimura §, Shinpei Kato †‡

† Dept. Computer Science, The University of Tokyo, Japan § Hitachi, Ltd., Japan ‡ Tier IV, Inc., Japan

2019/11/4

slide-2
SLIDE 2

1

In Autonomous Driving...

  • High Computational Tasks
  • Self-localization
  • Path Planning
  • Multiple Sensing Inputs
  • Location ：GPS / IMU
  • Scans ：Cameras / LiDAR
  • Map Data：Point-cloud Map / ADAS Map

Background

Map Data Scan Data Self-localization Path Planning

An Edge-Cloud System Model for Autonomous Vehicles (PPNIV '19) Yu Sasaki et al.

slide-3
SLIDE 3

2

Advance in embedded hardware

Automotive Embedded Systems

Pros Cons

  • Low power consumption
  • Small
  • Insufficient computation power
  • Low Upgradability

Commercial PC Embedded Hardware

An Edge-Cloud System Model for Autonomous Vehicles (PPNIV '19) Yu Sasaki et al.

slide-4
SLIDE 4

3

  • Creating ad-hoc network within vehicles
  • Communicating road information
  • Distributed computing
  • Information pass-thru
  • Cons
  • Not Ideal with a few vehicles
  • Interference from non-VANET vehicles

(Obstacle Shadowing) [Meireles+,2010]

Distributed Computing Approach

Vehicular Ad Hoc Networks (VANET) [Benslimane+,2004]

  • R. Meireles, et al: “Experimental Study on the Impact of Vehicular Obstructions in VANETs” , 2010
  • A. Benslimane et al: “Optimized Dissemination of Alarm Messages in Vehicular Ad-Hoc Networks (VANET)” , 2004

An Edge-Cloud System Model for Autonomous Vehicles (PPNIV '19) Yu Sasaki et al.

slide-5
SLIDE 5

4

In other fields...

Edge-Cloud Computing Approach

An Edge-Cloud System Model for Autonomous Vehicles (PPNIV '19) Yu Sasaki et al.

slide-6
SLIDE 6

5

  • Requires higher network

bandwidth

  • Advance in next-gen

wireless network

Why few edge-cloud model? Why now?

An Edge-Cloud System Model for Autonomous Vehicles (PPNIV '19) Yu Sasaki et al.

It has become more practical to consider edge-cloud model

slide-7
SLIDE 7

6

  • Edge-cloud system model for

autonomous vehicles

  • Effectiveness of presented

model

  • Round-trip execution time
  • Improvement rate
  • Case study using Autoware

[Kato+,2015]

Contribution

S Kato et al: “An Open Approach to Autonomous Vehicles”, 2015

An Edge-Cloud System Model for Autonomous Vehicles (PPNIV '19) Yu Sasaki et al.

slide-8
SLIDE 8

7

  • Execution time on edge machine
  • Execution time on cloud server
  • Communication cost between edge and cloud
  • Communication delay due to packet loss

Round-trip Execution Time

An Edge-Cloud System Model for Autonomous Vehicles (PPNIV '19) Yu Sasaki et al.

slide-9
SLIDE 9

8

Model Notation

  • Task :𝑈 = {𝑘%, … , 𝑘(}

Num of jobs :𝑂

  • Jobs run on cloud : 𝐾, Jobs run on edge : 𝐾-
  • Job pair with edge-cloud communication : 𝐹 = {(𝑞, 𝑟), … }
  • Execution time of 𝑘3 on cloud : 𝑓,,56 Execution time of on 𝑘3 edge : 𝑓-,56
  • Bandwidth : 𝐶

Packet loss : 𝑔

  • Traffic : 𝑢𝑠𝑏𝑜𝑡>,? average transmission speed : 𝑐>,?

System Model

𝑇 = ∑

56∈DE

𝑓,,56 + ∑

56∈DG

𝑓-,56 + (1 + 𝑔) ∑

(>,?)∈-

𝑢𝑠𝑏𝑜𝑡>,? 𝐶 ∑

(>,?)∈-

𝑐I,5 ≤ 𝐶 ∑

(>,?)∈-

𝑐I,5 > 𝐶 𝑇 = ∑

56∈DE

𝑓,,56 + ∑

56∈DG

𝑓-,56 + (1 + 𝑔) ∑

(>,?)∈-

𝑢𝑠𝑏𝑜𝑡>,? 𝑐>,?

An Edge-Cloud System Model for Autonomous Vehicles (PPNIV '19) Yu Sasaki et al.

slide-10
SLIDE 10

9

  • Experimented with wired Ethernet

connection (simulating 5G / Wireless 6 speed)

  • Constructed model assuming no

packet loss

Experiment

An Edge-Cloud System Model for Autonomous Vehicles (PPNIV '19) Yu Sasaki et al.

slide-11
SLIDE 11

10

  • Evaluated using simulation mode on Autoware [Kato+,2015]
  • ROS-based open-source software (https://www.autoware.ai/)
  • DRIVE PX2 as edge machine, desktop PC as cloud server
  • Evaluation
  • Self-localization
  • Round-trip execution time
  • Deadline miss rate
  • Path planner
  • Execution time
  • Each evaluation on
  • Edge (standalone)
  • Cloud(standalone)
  • Edge-cloud

Experiment

S Kato et al: “An Open Approach to Autonomous Vehicles”, 2015

An Edge-Cloud System Model for Autonomous Vehicles (PPNIV '19) Yu Sasaki et al.

slide-12
SLIDE 12

11

DRIVE PX2

An Edge-Cloud System Model for Autonomous Vehicles (PPNIV '19) Yu Sasaki et al.

https://devblogs.nvidia.com/wp-content/uploads/2018/04/self-driving-car-drive-px-autochauffeur-516-ud-1.png

  • 2 SoC board
  • Denver(x2)= Cortex-A57(x4)
  • Pascal GPU
  • 512 CUDA core
  • 4GB Mem
  • 6GB Memory
  • 64GB storage
  • 80W
slide-13
SLIDE 13

12

Self-localization

An Edge-Cloud System Model for Autonomous Vehicles (PPNIV '19) Yu Sasaki et al.

① /points_raw ② /filtered_points

① ②

slide-14
SLIDE 14

13

Grid Map Filter

An Edge-Cloud System Model for Autonomous Vehicles (PPNIV '19) Yu Sasaki et al.

③ /points_no_ground ① /points_raw ④ /realtimecostmap ⑤ /distance_transform

① ③ ④ ⑤

slide-15
SLIDE 15

14

Path Planning

An Edge-Cloud System Model for Autonomous Vehicles (PPNIV '19) Yu Sasaki et al.

⑤ ⑥ /current_pose

slide-16
SLIDE 16

15

Machine Environment Simulation Dataset

  • 200sec drive data acquired in Nagoya University
  • Deadline
  • Self-localization : 100ms
  • Path-planning : 2,000ms

Experiment Environment

Edge(DRIVE PX2) Cloud CPU Cortex-A57(x4), Denver(x2) Core i7-8700 CPU Freq. 2.0GHz(A57), 2.0GHz(Denver) 3.2GHz Memory 6GB 32GB GPU Memory 4GB 8GB(GTX1080) CUDA Core 512 1280 Linux kernel 4.9.38 4.16.5 ROS kinetic

An Edge-Cloud System Model for Autonomous Vehicles (PPNIV '19) Yu Sasaki et al.

slide-17
SLIDE 17

16

Simulation Data

An Edge-Cloud System Model for Autonomous Vehicles (PPNIV '19) Yu Sasaki et al.

slide-18
SLIDE 18

17

  • Measured execution time of self-localization for 2,000cycles (200sec)
  • Average: (Edge) 55.4ms-> (Edge-Cloud) 35.1ms
  • Improved 57%
  • WCET: (Edge) 348ms -> (Edge-Cloud) 103ms
  • Improved 3.5x

Results

An Edge-Cloud System Model for Autonomous Vehicles (PPNIV '19) Yu Sasaki et al.

slide-19
SLIDE 19

18

  • Measured 100 sets (each set consists of 2,000 cycles)
  • Total of 200,000 cycles
  • Worst deadline rates (Edge) 3%->(Edge-Cloud) 0.20%
  • Only 3 misses out of 200,000 cycles

Results

An Edge-Cloud System Model for Autonomous Vehicles (PPNIV '19) Yu Sasaki et al.

slide-20
SLIDE 20

19

  • Measured execution time of

path planner for 100 cycles

  • In some cases 10x faster than

standalone edge machine

  • Drastic difference in goals that

requires curving path

Results

An Edge-Cloud System Model for Autonomous Vehicles (PPNIV '19) Yu Sasaki et al.

slide-21
SLIDE 21

20

Result

An Edge-Cloud System Model for Autonomous Vehicles (PPNIV '19) Yu Sasaki et al.

slide-22
SLIDE 22

21

  • Measured execution time of

path planner for 100 cycles

  • In some cases 10x faster than

edge machine

  • Drastic difference in goals that

requires curving path

Results

An Edge-Cloud System Model for Autonomous Vehicles (PPNIV '19) Yu Sasaki et al.

slide-23
SLIDE 23

22

  • We presented edge-cloud computing model for autonomous

driving

  • Even with the communication cost, our model is faster and

more stable

  • Future Work

Tests in real world environment (5G / Wireless 6)

Conclusion

An Edge-Cloud System Model for Autonomous Vehicles (PPNIV '19) Yu Sasaki et al.

Questions?

E-Mail : Yu Sasaki (yu.sasaki@pf.is.s.u-tokyo.ac.jp)