Lightweight and Resource-Efficient Perception for Robotic Guide Dogs

2026-10-02

AI Summary

This study addresses the problem of erroneous accelerator placement decisions caused by resource contention during multi-camera streaming perception on heterogeneous edge platforms. We propose a novel evaluation paradigm based on contention scanning and the worst-stream sAP metric. Specifically, we construct an end-to-end multi-stream perception pipeline with hybrid GPU/NPU acceleration, revealing that isolated testing misleads deployment decisions, average sAP obscures severe per-stream degradation, and local GPU contention can invert optimal placement strategies. Experimental results demonstrate that under heavy workloads, the worst-stream sAP of the all-NPU configuration reaches 5.2 times that of the all-GPU setup, validating the robustness advantages of NPUs in concurrent scenarios.

Abstract

Multi-camera streaming perception is increasingly deployed on heterogeneous edge platforms shared with co-resident workloads, yet accelerator placement is often evaluated using isolated single-stream experiments and mean streaming average precision (sAP). Using two end-to-end pipelines on a single GPU--NPU platform, we show that isolated evaluation can mis-rank deployment-time placement. Although the GPU pipeline is preferred in isolation, GPU-localized contention introduces deadline misses that make detections stale and can reverse the preferred placement before full GPU saturation. The NPU pipeline is less accurate than the GPU pipeline on small and medium objects in isolation, but nearly matches it on large objects. The largest absolute sAP losses in our latency and contention experiments occur for large objects. In our four-stream experiments, the preferred placement depends on which path becomes stale, and increasing GPU-side contention shifts the best placement from All-GPU to All-NPU. Under a GPU-saturating vision--language co-tenant, All-NPU achieves $5.2\times$ the worst-stream sAP of All-GPU. Because mean sAP can hide severe single-stream degradation, evaluation should report contention sweeps, deadline-miss rates on both paths, and worst-stream sAP alongside mean sAP.

Problem

• edge computing
• multi-camera streaming perception
• resource contention
• accelerator placement
• heterogeneous platforms

Innovation

• heterogeneous edge platforms
• streaming perception
• accelerator placement
• worst-stream sAP
• resource contention