Human-Robot Interaction

Motion planning is a core problem in robotics, with a range of existing methods aimed to address its diverse set of challenges. However, most existing methods rely on complete knowledge of the robot environment; an assumption that seldom holds true due to inherent limitations of robot perception. To enable tractable motion planning for high-DOF robots under partial observability, we introduce BLIND, an algorithm that leverages human guidance. BLIND utilizes inverse reinforcement learning to derive motion-level guidance from human critiques. The algorithm overcomes the computational challenge of reward learning for high-DOF robots by projecting the robot’s continuous configuration space to a motion-planner-guided discrete task model. The learned reward is in turn used as guidance to generate robot motion using a novel motion planner. We demonstrate BLIND using the Fetch robot and perform two simulation experiments with partial observability. Our experiments demonstrate that, despite the challenge of partial observability and high dimensionality, BLIND is capable of generating safe robot motion and outperforms baselines on metrics of teaching efficiency, success rate, and path quality.

Communication is critical to collaboration; however, too much of it can degrade performance. Motivated by the need for effective use of a robot’s communication modalities, in this work, we present a computational framework that decides if, when, and what to communicate during human-robot collaboration. The framework, titled CommPlan, consists of a model specification process and an execution-time POMDP planner. To address the challenge of collecting interaction data, the model specification process is hybrid : where part of the model is learned from data, while the remainder is manually specified. Given the model, the robot’s decision-making is performed computationally during interaction and under partial observability of human’s mental states. We implement CommPlan for a shared workspace task, in which the robot has multiple communication options and needs to reason within a short time. Through experiments with human participants, we confirm that CommPlan results in the effective use of communication capabilities and improves human-robot collaboration.

Humans and machines often possess complementary skills. The recognition of this fact is leading to a steadily growing interest in collaborative robots. Despite the growing interest, however, a fundamental question remains to be answered: "How does one develop effective collaborative robots?" Three entities need to be considered while answering this question – namely, the collaborative robot itself, the human teammate whom the robot interacts with, and, equally importantly, the robot developer who is tasked with designing the machine. Each of these entities possesses different information. Effective sharing of this information is essential for developing collaborative robots and achieving fluent collaboration. In this dissertation, I present models and algorithms to enable effective information sharing between the robot, the human, and the developer. I begin by presenting the Agent Markov Model (AMM), a Bayesian model of sequential decision-making behavior, and Constrained Variational Inference (CVI), a hybrid learning algorithm that can learn generative models both from data and domain expertise. By utilizing AMM and CVI, the developer can specify decisionmaking models both for the human teammate and the collaborative robot with reduced labeling effort. Next, I present ADACORL, a framework to generate the collaborative robot’s policy for interaction. By leveraging algorithms for planning under uncertainty, ADACORL can generate fluent robot behavior for human-robot collaborative tasks with state spaces significantly larger than prior art (> 1 million states) and short planning times (< 1 s). Finally, I provide an approach for deciding if, when, and what to communicate during human-robot collaboration. Through human-robot interaction studies, I demonstrate that the proposed decision-making approaches result in the effective use of the robot’s action and communication capabilities during collaboration with a human teammate.

We consider human-robot collaboration in sequential tasks with known task objectives. For interaction planning in this setting, the utility of models for decision-making under uncertainty has been demonstrated across domains. However, in practice, specifying the model parameters remains challenging, requiring significant effort from the robot developer. To alleviate this challenge, we present ADACORL, a framework to specify decision-making models and generate robot behavior for interaction. Central to our approach are a factored task model and a semi-supervised algorithm to learn models of human behavior. We demonstrate that our specification approach, despite significantly fewer labels, generates models (and policies) that perform equally well or better than models learned with supervised data. By leveraging pre-computed performance bounds and an online planner, ADACORL can generate robot behavior for collaborative tasks with large state spaces (> 1 million states) and short planning times (< 0.5 s).

Artificial agents (both embodied robots and software agents) that interact with humans are increasing at an exceptional rate. Yet, achieving seamless collaboration between artificial agents and humans in the real world remains an active problem [Thomaz et al., 2016]. A key challenge is that the agents need to make decisions without complete information about their shared environment and collaborators. For instance, a human-robot team performing a rescue operation after a disaster may not have an accurate map of their surroundings. Even in structured domains, such as manufacturing, a robot might not know the goals or preferences of its human collaborators [Unhelkar et al., 2018]. Algorithmically, this challenge manifests itself as a problem of decision-making under uncertainty in which the agent has to reason about the latent states of its environment and human collaborator. However, in practice, quantifying this uncertainty (ie, the state transition function) and even specifying the features (ie, the relevant states) of human-machine collaboration is difficult. Thus, the objective of this thesis research is to develop novel algorithms that enable artificial agents to learn and reason about the latent states of humanmachine collaboration and achieve fluent interaction.

Robots that operate alongside or cooperatively with humans are envisioned as the next generation of robotics. Toward this vision, we present the first mobile robot system designed for and capable of operating on the moving floors of automotive final assembly lines (AFALs). AFALs represent a distinct challenge for mobile robots in the form of dynamic surfaces: the conveyor belts that transport cars throughout the factory during final assembly.

Introducing mobile robots into the collaborative assembly process poses unique challenges for ensuring efficient and safe human-robot interaction. Current human-robot work cells require the robot to cease operating completely whenever a human enters a shared region of the given cell, and the robots do not explicitly model or adapt to the behavior of the human. In this work, we present a human-aware robotic system with single-axis mobility that incorporates both predictions of human motion and planning in time to execute efficient and safe motions during automotive final assembly. We evaluate our system in simulation against three alternative methods, including a baseline approach emulating the behavior of standard safety systems in factories today. We also assess the system within a factory test environment. Through both live demonstration and results from simulated experiments, we show that our approach produces statistically significant improvements in quantitative measures of safety and fluency of interaction.

Communication between agents has the potential to improve team performance of collaborative tasks. However, communication is not free in most domains, requiring agents to reason about the costs and benefits of sharing information. In this work, we develop an online, decentralized communication policy, ConTaCT, that enables agents to decide whether or not to communicate during time-critical collaborative tasks in unknown, deterministic environments. Our approach is motivated by real-world applications, including the coordination of disaster response and search and rescue teams. These settings motivate a model structure that explicitly represents the world model as initially unknown but deterministic in nature, and that de-emphasizes uncertainty about action outcomes. Simulated experiments are conducted in which ConTaCT is compared to other multi-agent communication policies, and results indicate that ConTaCT achieves comparable task performance while substantially reducing communication overhead.

Mobile, interactive robots that operate in human-centric environments need the capability to safely and efficiently navigate around humans. This requires the ability to sense and predict human motion trajectories and to plan around them. In this paper, we present a study that supports the existence of statistically significant biomechanical turn indicators of human walking motions. Further, we demonstrate the effectiveness of these turn indicators as features in the prediction of human motion trajectories. Human motion capture data is collected with predefined goals to train and test a prediction algorithm. Use of anticipatory features results in improved performance of the prediction algorithm. Lastly, we demonstrate the closed-loop performance of the prediction algorithm using an existing algorithm for motion planning within dynamic environments. The anticipatory indicators of human walking motion can be used with different prediction and/or planning algorithms for robotics; the chosen planning and prediction algorithm demonstrates one such implementation for human-robot co-navigation.

Industrial robots are on the verge of emerging from their cages, and entering the final assembly to work along side humans. Towards this we are developing a collaborative robot capable of assisting humans in the final automotive assembly. Several algorithmic as well as design challenges exist when the robots enter the unpredictable, human-centric and time-critical environment of final assembly. In this work, we briefly discuss a few of these challenges along with developed solutions and proposed methodologies, and their implications for improving human-robot collaboration.

There is an emerging desire across manufacturing industries to deploy robots that support people in their manual work, rather than replace human workers. This paper explores one such opportunity, which is to field a mobile robotic assistant that travels between part carts and the automotive final assembly line, delivering tools and materials to the human workers. We compare the performance of a mobile robotic assistant to that of a human assistant to gain a better understanding of the factors that impact its effectiveness. Statistically significant differences emerge based on type of assistant, human or robot. Interaction times and idle times are statistically significantly higher for the robotic assistant than the human assistant. We report additional differences in participant’s subjective response regarding team fluency, situational awareness, comfort and safety. Finally, we discuss how results from the experiment inform the design of a more effective assistant.