Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Concurrent Active Learning in Autonomous Airborne Source Search: Dual Control for Exploration and Exploitation

Zhongguo Li, Wen-Hua Chen, Jun Yang

TL;DR

Compared with the information-theoretic approach, CL-DCEE not only guarantees convergence, but produces better search performance and consumes much less computational time.

Abstract

In this paper, a concurrent learning framework is developed for source search in an unknown environment using autonomous platforms equipped with onboard sensors. Distinct from the existing solutions that require significant computational power for Bayesian estimation and path planning, the proposed solution is computationally affordable for onboard processors. A new concept of concurrent learning using multiple parallel estimators is proposed to learn the operational environment and quantify estimation uncertainty. The search agent is empowered with dual capability of exploiting current estimated parameters to track the source and probing the environment to reduce the impacts of uncertainty, namely Concurrent Learning based Dual Control for Exploration and Exploitation (CL-DCEE). In this setting, the control action not only minimises the tracking error between future agent's position and estimated source location, but also the uncertainty of predicted estimation. More importantly, the rigorous proven properties such as the convergence of CL-DCEE algorithm are established under mild assumptions on noises, and the impact of noises on the search performance is examined. Simulation results are provided to validate the effectiveness of the proposed CL-DCEE algorithm. Compared with the information-theoretic approach, CL-DCEE not only guarantees convergence, but produces better search performance and consumes much less computational time.

Concurrent Active Learning in Autonomous Airborne Source Search: Dual Control for Exploration and Exploitation

TL;DR

Compared with the information-theoretic approach, CL-DCEE not only guarantees convergence, but produces better search performance and consumes much less computational time.

Abstract

In this paper, a concurrent learning framework is developed for source search in an unknown environment using autonomous platforms equipped with onboard sensors. Distinct from the existing solutions that require significant computational power for Bayesian estimation and path planning, the proposed solution is computationally affordable for onboard processors. A new concept of concurrent learning using multiple parallel estimators is proposed to learn the operational environment and quantify estimation uncertainty. The search agent is empowered with dual capability of exploiting current estimated parameters to track the source and probing the environment to reduce the impacts of uncertainty, namely Concurrent Learning based Dual Control for Exploration and Exploitation (CL-DCEE). In this setting, the control action not only minimises the tracking error between future agent's position and estimated source location, but also the uncertainty of predicted estimation. More importantly, the rigorous proven properties such as the convergence of CL-DCEE algorithm are established under mild assumptions on noises, and the impact of noises on the search performance is examined. Simulation results are provided to validate the effectiveness of the proposed CL-DCEE algorithm. Compared with the information-theoretic approach, CL-DCEE not only guarantees convergence, but produces better search performance and consumes much less computational time.

Paper Structure

This paper contains 14 sections, 42 equations, 8 figures, 3 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. Problem Formulation
  3. Agent Modelling
  4. Dispersion Modelling
  5. Sensor Modelling
  6. Objective Function Construction
  7. Concurrent Learning for Dual Control with Exploration and Exploitation
  8. Framework of Concurrent Learning with Dual Effects
  9. Convergence Analysis
  10. Simulation Study
  11. Unbiased Sensor Noises in Partially Unknown Environment
  12. Unknown Environment with Sensor Non-detection Events
  13. Experimental Study
  14. Conclusion

Figures (8)

  • Figure 1: Mean-square-error between the estimated and true source positions.
  • Figure 2: Distance between agent's position and the true source.
  • Figure 3: Representative runs of different algorithms: (a) CL-DCEE with $N=10$, (b) CL-DCEE with $N=100$, (c) Entrotaxis. Red lines are the paths of the UAV, the green dots represent the estimated source position, and the black dots represent the true source position.
  • Figure 4: Performance of CL-DCEE algorithm with different number of estimators.
  • Figure 5: Mean-square-error between estimated and true source positions with unknown environment and sensor dropouts.
  • ...and 3 more figures