Real-time Pedestrian Detection Approach with an Efficient Data Communication Bandwidth Strategy

TL;DR

This paper presents a real-time pedestrian detection system using deep learning and lossy data compression to significantly reduce bandwidth needs while maintaining high detection accuracy at signalized intersections.

Contribution

It introduces an edge computing approach that combines pedestrian detection with an efficient data compression technique to optimize bandwidth use without sacrificing accuracy.

Findings

Achieved 31x reduction in bandwidth from 9.82 Mbits/sec to 0.31 Mbits/sec.

Maintained pedestrian detection accuracy at PSNR of 43 dB.

Demonstrated effective tradeoff between data compression and detection performance.

Abstract

Vehicle-to-Pedestrian (V2P) communication can significantly improve pedestrian safety at a signalized intersection. It is unlikely that pedestrians will carry a low latency communication enabled device and activate a pedestrian safety application in their hand-held device all the time. Because of this limitation, multiple traffic cameras at the signalized intersection can be used to accurately detect and locate pedestrians using deep learning and broadcast safety alerts related to pedestrians to warn connected and automated vehicles around a signalized intersection. However, unavailability of high-performance computing infrastructure at the roadside and limited network bandwidth between traffic cameras and the computing infrastructure limits the ability of real-time data streaming and processing for pedestrian detection. In this paper, we develop an edge computing based real-time pedestrian detection strategy combining pedestrian detection algorithm using deep learning and an efficient data communication approach to reduce bandwidth requirements while maintaining a high object detection accuracy. We utilize a lossy compression technique on traffic camera data to determine the tradeoff between the reduction of the communication bandwidth requirements and a defined object detection accuracy. The performance of the pedestrian-detection strategy is measured in terms of pedestrian classification accuracy with varying peak signal-to-noise ratios. The analyses reveal that we detect pedestrians by maintaining a defined detection accuracy with a peak signal-to-noise ratio (PSNR) 43 dB while reducing the communication bandwidth from 9.82 Mbits/sec to 0.31 Mbits/sec, a 31x reduction.