Hardware Acceleration of Robot Scene Perception Algorithms

Hybrid machine learning algorithms that combine deep learning with probabilistic inference techniques provide highly accurate scene perception for robot manipulation. In particular, a 2-stage approach that combines object detection using convolutional neural networks with Monte-Carlo sampling for po...

Full description

Saved in:
Bibliographic Details
Published in:Digest of technical papers - IEEE/ACM International Conference on Computer-Aided Design pp. 1 - 8
Main Authors: Liu, Yanqi, Derman, Can Eren, Calderoni, Giuseppe, Bahar, R. Iris
Format: Conference Proceeding
Language:English
Published: Association on Computer Machinery 02.11.2020
Subjects:
2-stage approach
Acceleration
accurate energy-efficient scene perception
adversarial scenarios
complexity-aware techniques
computational complexity
convolutional neural networks
deep learning
energy consumption
Feature extraction
Hardware
high computational complexity
high-end GPU
highly accurate scene perception
hybrid machine
Image edge detection
image sampling
inference mechanisms
learning (artificial intelligence)
low-power
low-power GPU implementation
Monte Carlo methods
Monte-Carlo sampling
neural nets
Neural networks
novel hardware accelerator
object detection
pose estimation
probabilistic inference techniques
resource utilization
robot manipulation
robot scene perception algorithms
robotics
Robots
robust robot perception
targeted hardware acceleration
Three-dimensional displays
ISSN:1558-2434
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Hybrid machine learning algorithms that combine deep learning with probabilistic inference techniques provide highly accurate scene perception for robot manipulation. In particular, a 2-stage approach that combines object detection using convolutional neural networks with Monte-Carlo sampling for pose estimation has been shown to perform particularly well under adversarial scenarios. Unfortunately, this accuracy comes at the cost of high computational complexity, which affects runtime, resource utilization, and energy consumption. This paper describes various challenges in developing complexity-aware techniques for robust robot perception and presents a novel hardware accelerator that addresses these challenge. Experimental results show our design is at least 30% faster and consumes 97% less energy compared to an implementation on a high-end GPU. Compared to a low-power GPU implementation, our design is 95% faster while consuming 96% less energy, demonstrating that accurate, energy-efficient scene perception is possible in real time with targeted hardware acceleration.
ISSN:1558-2434
DOI:10.1145/3400302.3415766