Wearable Multi-Sensor and HMM Enabled Stress State Evaluation In Vivo Fish During Waterless Transportation Process

Vivo fish stress monitoring and evaluation have become a significant means to improve fish health and quality during waterless transportation process. Wearable sensor monitoring has attracted great interest in fish stress state detection. In this paper, we presented wearable multi-sensor based monitoring system to obtain the critical stress data and HMM-based coupling modeling to detect vivo fish stress state during waterless transportation process. The wearable temperature sensor, wearable blood glucose sensor and wearable breath sensor were designed and implemented, which can continuously capture the data for further analysis and track to key stress parameters. A novel data-driven approach using particle swarm optimization (PSO) - Hidden Markov Model (HMM) algorithm was proposed to realize fish stress state evaluation in the current and next period during waterless transportation process. The PSO-HMM based model accuracy for different stress state evaluation reached 90.43%, 93.23%,92.58%, respectively. This work contributes to improve vivo fish health and reduce quality loss during waterless transportation process.