Deep reinforcement learning has emerged as a popular and powerful way to develop locomotion controllers for quadruped robots. Common approaches have largely focused on learning actions directly in joint space, or learning to modify and offset foot positions produced by trajectory generators. Both approaches typically require careful reward shaping and training for millions of time steps, and with trajectory generators introduce human bias into the resulting control policies. In this paper, we present a learning framework that leads to the natural emergence of fast and robust bounding policies for quadruped robots. The agent both selects and controls actions directly in task space to track desired velocity commands subject to environmental noise including model uncertainty and rough terrain. We observe that this framework improves sample efficiency, necessitates little reward shaping, leads to the emergence of natural gaits such as galloping and bounding, and eases the sim-to-real transfer at running speeds. Policies can be learned in only a few million time steps, even for challenging tasks of running over rough terrain with loads of over 100% of the nominal quadruped mass. Training occurs in PyBullet, and we perform a sim-to-sim transfer to Gazebo and sim-to-real transfer to the Unitree A1 hardware. For sim-to-sim, our results show the quadruped is able to run at over 4 m/s without a load, and 3.5 m/s with a 10 kg load, which is over 83% of the nominal quadruped mass. For sim-to-real, the Unitree A1 is able to bound at 2 m/s with a 5 kg load, representing 42% of the nominal quadruped mass.

Developing intelligent energy management systems with high adaptability and superiority is necessary and significant for Hybrid Electric Vehicles (HEVs). This paper proposed an ensemble learning-based scheme based on a learning automata module (LAM) to enhance vehicle energy efficiency. Two parallel base learners following two exploration-to-exploitation ratios (E2E) methods are used to generate an optimal solution, and the final action is jointly determined by the LAM using three ensemble methods. 'Reciprocal function-based decay' (RBD) and 'Step-based decay' (SBD) are proposed respectively to generate E2E ratio trajectories based on conventional Exponential decay (EXD) functions of reinforcement learning. Furthermore, considering the different performances of three decay functions, an optimal combination with the RBD, SBD, and EXD is employed to determine the ultimate action. Experiments are carried out in software-in-loop (SiL) and hardware-in-the-loop (HiL) to validate the potential performance of energy-saving under four predefined cycles. The SiL test demonstrates that the ensemble learning system with an optimal combination can achieve 1.09$\%$ higher vehicle energy efficiency than a single Q-learning strategy with the EXD function. In the HiL test, the ensemble learning system with an optimal combination can save more than 1.04$\%$ in the predefined real-world driving condition than the single Q-learning scheme based on the EXD function.

Reinforcement learning (RL) has become widely adopted in robot control. Despite many successes, one major persisting problem can be very low data efficiency. One solution is interactive feedback, which has been shown to speed up RL considerably. As a result, there is an abundance of different strategies, which are, however, primarily tested on discrete grid-world and small scale optimal control scenarios. In the literature, there is no consensus about which feedback frequency is optimal or at which time the feedback is most beneficial. To resolve these discrepancies we isolate and quantify the effect of feedback frequency in robotic tasks with continuous state and action spaces. The experiments encompass inverse kinematics learning for robotic manipulator arms of different complexity. We show that seemingly contradictory reported phenomena occur at different complexity levels. Furthermore, our results suggest that no single ideal feedback frequency exists. Rather that feedback frequency should be changed as the agent's proficiency in the task increases.

Unmanned aerial vehicles (UAVs) are seen as a promising technology to perform a wide range of tasks in wireless communication networks. In this work, we consider the deployment of a group of UAVs to collect the data generated by IoT devices. Specifically, we focus on the case where the collected data is time-sensitive, and it is critical to maintain its timeliness. Our objective is to optimally design the UAVs' trajectories and the subsets of visited IoT devices such as the global Age-of-Updates (AoU) is minimized. To this end, we formulate the studied problem as a mixed-integer nonlinear programming (MINLP) under time and quality of service constraints. To efficiently solve the resulting optimization problem, we investigate the cooperative Multi-Agent Reinforcement Learning (MARL) framework and propose an RL approach based on the popular on-policy Reinforcement Learning (RL) algorithm: Policy Proximal Optimization (PPO). Our approach leverages the centralized training decentralized execution (CTDE) framework where the UAVs learn their optimal policies while training a centralized value function. Our simulation results show that the proposed MAPPO approach reduces the global AoU by at least a factor of 1/2 compared to conventional off-policy reinforcement learning approaches.

When applying reinforcement learning (RL) to a new problem, reward engineering is a necessary, but often difficult and error-prone task a system designer has to face. To avoid this step, we propose LR4GPM, a novel (deep) RL method that can optimize a global performance metric, which is supposed to be available as part of the problem description. LR4GPM alternates between two phases: (1) learning a (possibly vector) reward function used to fit the performance metric, and (2) training a policy to optimize an approximation of this performance metric based on the learned rewards. Such RL training is not straightforward since both the reward function and the policy are trained using non-stationary data. To overcome this issue, we propose several training tricks. We demonstrate the efficiency of LR4GPM on several domains. Notably, LR4GPM outperforms the winner of a recent autonomous driving competition organized at DAI'2020.

A prominent approach to visual Reinforcement Learning (RL) is to learn an internal state representation using self-supervised methods, which has the potential benefit of improved sample-efficiency and generalization through additional learning signal and inductive biases. However, while the real world is inherently 3D, prior efforts have largely been focused on leveraging 2D computer vision techniques as auxiliary self-supervision. In this work, we present a unified framework for self-supervised learning of 3D representations for motor control. Our proposed framework consists of two phases: a pretraining phase where a deep voxel-based 3D autoencoder is pretrained on a large object-centric dataset, and a finetuning phase where the representation is jointly finetuned together with RL on in-domain data. We empirically show that our method enjoys improved sample efficiency in simulated manipulation tasks compared to 2D representation learning methods. Additionally, our learned policies transfer zero-shot to a real robot setup with only approximate geometric correspondence, and successfully solve motor control tasks that involve grasping and lifting from a single, uncalibrated RGB camera. Code and videos are available at https://yanjieze.com/3d4rl/ .

In Reinforcement Learning the Q-learning algorithm provably converges to the optimal solution. However, as others have demonstrated, Q-learning can also overestimate the values and thereby spend too long exploring unhelpful states. Double Q-learning is a provably convergent alternative that mitigates some of the overestimation issues, though sometimes at the expense of slower convergence. We introduce an alternative algorithm that replaces the max operation with an average, resulting also in a provably convergent off-policy algorithm which can mitigate overestimation yet retain similar convergence as standard Q-learning.

Deep brain stimulation (DBS) has shown great promise toward treating motor symptoms caused by Parkinson's disease (PD), by delivering electrical pulses to the Basal Ganglia (BG) region of the brain. However, DBS devices approved by the U.S. Food and Drug Administration (FDA) can only deliver continuous DBS (cDBS) stimuli at a fixed amplitude; this energy inefficient operation reduces battery lifetime of the device, cannot adapt treatment dynamically for activity, and may cause significant side-effects (e.g., gait impairment). In this work, we introduce an offline reinforcement learning (RL) framework, allowing the use of past clinical data to train an RL policy to adjust the stimulation amplitude in real time, with the goal of reducing energy use while maintaining the same level of treatment (i.e., control) efficacy as cDBS. Moreover, clinical protocols require the safety and performance of such RL controllers to be demonstrated ahead of deployments in patients. Thus, we also introduce an offline policy evaluation (OPE) method to estimate the performance of RL policies using historical data, before deploying them on patients. We evaluated our framework on four PD patients equipped with the RC+S DBS system, employing the RL controllers during monthly clinical visits, with the overall control efficacy evaluated by severity of symptoms (i.e., bradykinesia and tremor), changes in PD biomakers (i.e., local field potentials), and patient ratings. The results from clinical experiments show that our RL-based controller maintains the same level of control efficacy as cDBS, but with significantly reduced stimulation energy. Further, the OPE method is shown effective in accurately estimating and ranking the expected returns of RL controllers.

Sequential decision making in the real world often requires finding a good balance of conflicting objectives. In general, there exist a plethora of Pareto-optimal policies that embody different patterns of compromises between objectives, and it is technically challenging to obtain them exhaustively using deep neural networks. In this work, we propose a novel multi-objective reinforcement learning (MORL) algorithm that trains a single neural network via policy gradient to approximately obtain the entire Pareto set in a single run of training, without relying on linear scalarization of objectives. The proposed method works in both continuous and discrete action spaces with no design change of the policy network. Numerical experiments in benchmark environments demonstrate the practicality and efficacy of our approach in comparison to standard MORL baselines.

The challenges of using inadequate online recruitment systems can be addressed with machine learning and software engineering techniques. Bi-directional personalization reinforcement learning-based architecture with active learning can get recruiters to recommend qualified applicants and also enable applicants to receive personalized job recommendations. This paper focuses on how machine learning techniques can enhance the recruitment process in the travel nursing industry by helping speed up data acquisition using a multi-model data service and then providing personalized recommendations using bi-directional reinforcement learning with active learning. This need was especially evident when trying to respond to the overwhelming needs of healthcare facilities during the COVID-19 pandemic. The need for traveling nurses and other healthcare professionals was more evident during the lockdown period. A data service was architected for job feed processing using an orchestration of natural language processing (NLP) models that synthesize job-related data into a database efficiently and accurately. The multi-model data service provided the data necessary to develop a bi-directional personalization system using reinforcement learning with active learning that could recommend travel nurses and healthcare professionals to recruiters and provide job recommendations to applicants using an internally developed smart match score as a basis. The bi-directional personalization reinforcement learning-based architecture with active learning combines two personalization systems - one that runs forward to recommend qualified candidates for jobs and another that runs backward and recommends jobs for applicants.