| description abstract | This article focuses on improving the speed, accuracy, and robustness of autonomous aerial-based chemical sensing for plume mapping and source localization through characterizing, modeling, and feedforward compensation of gas-sensor dynamics. First, the dynamics of three types of gas sensors are modeled. Second, the maximum chemical-mapping speed is calculated and shown to be inversely proportional to sensor time constant. Third, an inversion-based approach is used to compensate for the sensor dynamics to improve mapping throughput. Results show that dynamics compensation enhances the chemical-mapping speed by over five times compared to the uncompensated case. Finally, to further demonstrate utility, the approach is applied to a particle swarm optimization example for plume-source localization. The improvement is observed by how well the agents converge to the true chemical gas source location when gas-sensor dynamics are taken into account. Specifically, for a static Gaussian plume source, feedforward compensation leads to 64% average improvement in localization success, and for a dynamic Quick Urban and Industrial Complex (QUIC) dispersion plume source, a 39% average improvement is observed. These results underscore the importance of sensor dynamics compensation for enhancing mapping and source localization throughput, accuracy, and robustness. | |
