CADENCE: Context-Adaptive Depth Estimation for Navigation and Computational Efficiency

TL;DR

CADENCE is an adaptive depth estimation system for autonomous navigation that dynamically adjusts computational effort based on environmental context, significantly reducing power and latency while improving accuracy.

Contribution

It introduces a novel context-adaptive system for depth estimation that optimizes computational resources in autonomous vehicles.

Findings

Reduces sensor acquisitions by 9.67%

Decreases inference latency by 74.8%

Improves navigation accuracy by 7.43%

Abstract

Autonomous vehicles deployed in remote environments typically rely on embedded processors, compact batteries, and lightweight sensors. These hardware limitations conflict with the need to derive robust representations of the environment, which often requires executing computationally intensive deep neural networks for perception. To address this challenge, we present CADENCE, an adaptive system that dynamically scales the computational complexity of a slimmable monocular depth estimation network in response to navigation needs and environmental context. By closing the loop between perception fidelity and actuation requirements, CADENCE ensures high-precision computing is only used when mission-critical. We conduct evaluations on our released open-source testbed that integrates Microsoft AirSim with an NVIDIA Jetson Orin Nano. As compared to a state-of-the-art static approach, CADENCE decreases sensor acquisitions, power consumption, and inference latency by 9.67%, 16.1%, and 74.8%, respectively. The results demonstrate an overall reduction in energy expenditure by 75.0%, along with an increase in navigation accuracy by 7.43%.