This work presents advances in predictive modeling of weed growth, as well as an improved planning index to be used in conjunction with these techniques, for the purpose of improving the performance of coordinated wee...
详细信息
This work presents advances in predictive modeling of weed growth, as well as an improved planning index to be used in conjunction with these techniques, for the purpose of improving the performance of coordinated weeding algorithms being developed for industrial agriculture. We demonstrate that the evolving Gaussian process (E-GP) method applied to measurements from the agents can predict the evolution of the field within the realistic simulation environment, Weed World. This method also provides physical insight into the seed bank distribution of the field. In this work, we extend the E-GP model in two important ways. First, we have developed a model that has a bias term, and we show how it is connected to the seed bank distribution. Second, we show that one may decouple the component of the model representing weed growth from the component, which varies with the seed bank distribution, and adapt the latter online. We compare this predictive approach with one that relies on known properties of the weed growth model and show that the E-GP method can drive down the total weed biomass for fields with high seed bank densities using less agents, without assuming this model information. We use an improved planning index, the Whittle index, which allows a balanced tradeoff between exploiting a row or allowing it to accrue reward and conforms to what we show is the theoretical limit for the fewest number of agents, which can be used in this domain.
This dissertation formulates a signed real measure of sublanguages of regular languages based on the principle of automata theory andreal analysis. The measure allows total ordering of any set of partially ordered sub...
详细信息
This dissertation formulates a signed real measure of sublanguages of regular languages based on the principle of automata theory andreal analysis. The measure allows total ordering of any set of partially ordered sublanguages of a regular language for quantitative evaluation of the controlled behavior of deterministic finite state automata (DFSA) under differentsupervisors. The computational complexity of the language measure algorithm is polynomial in the number of DFSA states. An online parameter identification procedure is presented for computation of the language measure parameters.A discrete event behavior-based multi-robot system has been designed and constructed to validate the language measure theory and its applications to supervisory control in the discrete-event setting. Each robot is equipped with multiple sensors and multiple actuators. The interactions between the robot(s) and the (possibly) dynamically changing environment are characterized by discrete-event and continuous models, and the design and analysis of the robotic system are presented in both continuous-time and discrete event domains. The robustness and reliability of the controlled behavior is guaranteed in the continuous-time domain. For example, visual servoing is applied to robot navigation during `emph{approaching target}' and the vector field historgam (VFH) method is used for robust `emph{obstacle avoidance}'. The discrete event interactions between behaviors are formulated as a supervisory control theory problem, where multiple supervisors are synthesized and implemented online for robot control under different specifications. The efficacy of the language parameter identification procedure is demonstratedin real-time supervisory control through experiments on the mobile robotic system as well as on a high-fidelity robot simulator. A quantitative performance measure has been used toevaluate various discrete event supervisory (DES) controllers and is validated through experiments. The
暂无评论