Comparison and Improvement of Bioinspired Mobile Algorithms to Trace the Emission Source Based on the Simulation Scenarios

Hazardous gas emissions may have serious consequences for surrounding residents and the environment. Bioinspired mobile robots equipped with gas sensors have the potential to become a solution for precisely tracking and locating emission sources. In this study, the performance, efficiency, and accuracy of various bionic algorithms with bioinspired mobile sensors, i.e., silkworm, E. coli, ZigZag, and step-up algorithms, were compared using field simulations to track emission sources in the atmosphere. In the tracing process, the determination criteria of maximum concentration, minimum concentration (i.e., 0), and concentration gradient were discussed quantitatively. The simulation results showed that the silkworm algorithm has the best performance in locating the emission source, while the E. coli algorithm has the highest tracking efficiency. Therefore, a single source-determination criterion may be insufficient, since tracking accuracy and efficiency can vary with different simulation algorithms. To address these concerns, a new tracking strategy driven by the inverse motion and interface gradient (RMIG) was proposed, based on the behaviors of female mosquitoes seeking hosts by tracking CO2 plumes, to improve tracking efficiency. It was found that the locating efficiency driven by RMIG is greatly improved and higher than that of the E. coli algorithm in the tested cases, with 40% to 100% explicitly enhanced. Finally, the optimal correlated matching of concentration distribution (OCMCD) method was used to locate the source with a mobile sensor. Compared with traditional and common source-determination criteria, the RMIG-OCMCD method can significantly improve location accuracy. The proposed RMIG-OCMCD method could be a practical choice for tracking emission sources in the atmosphere if an appropriate search strategy is designed.