Decentralized Trust in Crowdsourced Drone Services

Abstract

The proliferation of crowdsourced drone services in areas such as bushfire monitoring, environmental surveillance, and emergency response highlights the urgent need for scalable, secure, and reliable trust management systems. This thesis focuses on establishing and maintaining decentralized trust as drones are increasingly used for distributed crowdsourced services. The D2XChain framework implements a decentralized blockchain-based PKI system that secures drone communication by distributing the CA across multiple blockchains and eliminating single points of failure. The DDRM framework uses blockchain to promote transparency and integrity in drone service evaluations. It also employs a dual-token system to mitigate reputation manipulation and enhance operational trust. The Privacy-Preserving Framework leverages local differential privacy to anonymize data and blockchain technology to securely share it. It strikes a balance between ensuring functionality for efficient crowdsourced drone operations and preserving individual privacy. The DroneSSL framework uses spatial crowdsourcing and TinyML for real-time, scalable anomaly detection in resource-limited environments. This improves the timeliness and specificity of emergency responses and remote environmental monitoring. The UAVGuard framework prevents label-flipping attacks on GNNs in spatial crowdsourcing networks. Through self-supervised learning and community detection, UAVGuard ensures the integrity of trust assessments in drone services. The GALTrust framework uses GANs and type-2 fuzzy logic to provide adaptive and resilient trust evaluations, particularly for drone services in adversarial settings like bushfire monitoring. The TMIoDT framework combines digital twin technology with blockchain for real-time trust assessments. By replicating drone environments in virtualized spaces, it ensures secure and accurate decentralized trust management.