VLM Can Be a Good Assistant: Enhancing Embodied Visual Tracking with Self-Improving Vision-Language Models

TL;DR

This paper presents a self-improving framework that integrates vision-language models with active visual tracking to enhance failure recovery and significantly improve success rates in dynamic environments.

Contribution

It introduces the first integration of VLM-based reasoning for proactive failure recovery in embodied visual tracking systems, with a memory-augmented self-reflection mechanism.

Findings

Boosts success rates by 72% with state-of-the-art RL-based methods.

Achieves 220% improvement with PID-based methods.

Demonstrates significant performance gains in challenging environments.

Abstract

We introduce a novel self-improving framework that enhances Embodied Visual Tracking (EVT) with Vision-Language Models (VLMs) to address the limitations of current active visual tracking systems in recovering from tracking failure. Our approach combines the off-the-shelf active tracking methods with VLMs' reasoning capabilities, deploying a fast visual policy for normal tracking and activating VLM reasoning only upon failure detection. The framework features a memory-augmented self-reflection mechanism that enables the VLM to progressively improve by learning from past experiences, effectively addressing VLMs' limitations in 3D spatial reasoning. Experimental results demonstrate significant performance improvements, with our framework boosting success rates by $72\%$ with state-of-the-art RL-based approaches and $220\%$ with PID-based methods in challenging environments. This work represents the first integration of VLM-based reasoning to assist EVT agents in proactive failure recovery, offering substantial advances for real-world robotic applications that require continuous target monitoring in dynamic, unstructured environments. Project website: https://sites.google.com/view/evt-recovery-assistant.