Secure and Trustworthy Cooperation Framework for Multi-Vendor Autonomous Systems

Abstract

The proliferation of multi-vendor autonomous systems (MVAS), such as shared warehouses, necessitates robust cooperation frameworks for robotic agents from different vendors. Ensuring operational integrity in these environments is a critical challenge, as the misbehaviour of a single agent can disrupt the entire system. This paper proposes a decentralised framework for secure and trustworthy multi-robot cooperation. Our approach leverages distributed ledger technology (DLT) to create an immutable record of operations and introduces a novel, lightweight Proof of Location (PoL) mechanism. This PoL allows agents to verify task completions both physically and cryptographically with minimal computational overhead. The framework is further strengthened by a real-time, algorithmic trust and reputation system that continuously evaluates agent behaviour. A key advantage over traditional methods is our support for asynchronous verification, which enhances both scalability and efficiency by removing centralised bottlenecks. Extensive simulations in a warehouse setting demonstrate that our framework effectively detects and mitigates malicious activities, thereby improving the security and performance of MVAS. The broader applicability of our approach is further illustrated through a conceptual case study in a smart city parking scenario.