Robustifying the Deployment of tinyML Models for Autonomous mini-vehicles

TL;DR

This paper presents a robust, low-power tinyML approach for autonomous mini-vehicles that learns in real-time from the environment, improving robustness and efficiency using specialized hardware.

Contribution

It introduces a closed-loop learning framework with tinyCNNs and demonstrates a highly efficient implementation on ultra-low-power hardware for autonomous mini-vehicles.

Findings

GAP8 outperforms STM32L4 and NXP k64f in latency and energy consumption.

TinyCNNs improve robustness to lighting conditions.

Real-time learning enhances autonomous mini-vehicle performance.

Abstract

Standard-size autonomous navigation vehicles have rapidly improved thanks to the breakthroughs of deep learning. However, scaling autonomous driving to low-power systems deployed on dynamic environments poses several challenges that prevent their adoption. To address them, we propose a closed-loop learning flow for autonomous driving mini-vehicles that includes the target environment in-the-loop. We leverage a family of compact and high-throughput tinyCNNs to control the mini-vehicle, which learn in the target environment by imitating a computer vision algorithm, i.e., the expert. Thus, the tinyCNNs, having only access to an on-board fast-rate linear camera, gain robustness to lighting conditions and improve over time. Further, we leverage GAP8, a parallel ultra-low-power RISC-V SoC, to meet the inference requirements. When running the family of CNNs, our GAP8's solution outperforms any other implementation on the STM32L4 and NXP k64f (Cortex-M4), reducing the latency by over 13x and the energy consummation by 92%.