Search for: All records

Gradient-based methods have been widely used for system design and optimization in diverse application domains. Recently, there has been a renewed interest in studying theoretical properties of these methods in the context of control and reinforcement learning. This article surveys some of the recent developments on policy optimization, a gradient-based iterative approach for feedback control synthesis that has been popularized by successes of reinforcement learning. We take an interdisciplinary perspective in our exposition that connects control theory, reinforcement learning, and large-scale optimization. We review a number of recently developed theoretical results on the optimization landscape, global convergence, and sample complexityof gradient-based methods for various continuous control problems, such as the linear quadratic regulator (LQR), [Formula: see text] control, risk-sensitive control, linear quadratic Gaussian (LQG) control, and output feedback synthesis. In conjunction with these optimization results, we also discuss how direct policy optimization handles stability and robustness concerns in learning-based control, two main desiderata in control engineering. We conclude the survey by pointing out several challenges and opportunities at the intersection of learning and control. 

Adaptive task allocation is used in many human-machine systems and has been proven to improve operators’ performance with automated systems. However, there has been limited knowledge surrounding the benefits of adaptive task allocation in automated vehicles. In this study, participants were presented with photos and videos depicting driving scenarios of low or high workloads at two levels of automation (SAE Levels 2 and 3). The participants reported which tasks they felt comfortable allocating to themselves or to the driving automation system (DAS) in each driving scenario, as well as whether they would conduct the task allocation manually or have the DAS automatically allocate the tasks. Our results showed that participants preferred conducting manual task allocation and preferred the system to complete more tasks when the perceived workload was high. There was no significant difference between the high and low workload scenarios in terms of whether participants chose to allocate tasks.