Design and Implementation of a Voice Controlled Indoor Autonomous Robot Kit

Ren Zhong, Zhaofeng Tian, Qiren Wang, Mingyu Guo, Weisong Shi

Abstract: The integration of voice interaction and autonomous driving into robotic systems holds immense potential for enhancing mobility aids in environments such as homes and healthcare facilities. Despite advancements in voice technology and navigation systems, existing indoor robots and mobility aids often lack intuitive, secure, and adaptive functionalities. This paper introduces VOCAR (Voice-Controlled Autonomous Robot Kit), a modular system designed to enable natural voice-based interaction and efficient autonomous navigation for mobile robots. VOCAR features personalized voiceprint authentication, a large language model (LLM)-based intent recognition pipeline, and advanced planning modules for obstacle avoidance, narrow passage traversal, and energy-efficient trajectory generation. It employs a unified CAN bus protocol to ensure compatibility with diverse robotic platforms. The feasibility of VOCAR is demonstrated through its implementation on a powered wheelchair, showcasing its ability to enhance user autonomy and interaction in dynamic indoor environments. This work highlights VOCAR's potential to transform mobility aids by providing a secure, adaptive, and energy-conscious solution for voice-controlled autonomous navigation.