Multi-Sensors-Based Physiological Stress Monitoring and Online Survival Prediction System for Live Fish Waterless Transportation

Live fish waterless transport strategy for the purpose of consumption or ornament is a novel technology, which can implement the larger volume transportation with high survival results as well as less wastewater pollution. The aim of this study was to establish an accurate survival prediction system based on the least physiological stress variations in different temperature ranges, which statistically analyzes the trend of key stress factors under the well life-supported temperatures to improve the final survival result. Furthermore, the accuracy of survival prediction can be adaptively and dynamically improved in consideration of the historical survival series in different seasons. In order to verify the practicability and performance of the system, Paralichthys olivaceus was selected as experimental subjects to carry out dynamic survival prediction experiments for over 30 hours waterless transportation, and the accuracy of the survival prediction system has verified over 97.2% in 0.5-2.5C°, which show that mean absolute error (MAE) and mean squared error (MSE) by less than 0.2 and 0.07 to produce the most accurate estimation in the comparison. The development of this survival prediction system can significantly optimize the current waterless delivery management by reducing the potential mortality and providing the managerial references for industrialization of the live fish waterless logistics.