Online Trajectory Planning Algorithms For Robotic Systems Under Uncertainty In Interactive Environments

The mission of this thesis is to develop algorithms for planning and control of intelligent mobile robots that operate autonomously in open, interactive environments. Presence of other agents and objects in such an environment makes planning significantly challenging, as they inevitably bring about environmental and dynamic uncertainty that the robot must properly handle. Despite recent advances in perception, planning and control, many existing robotic systems to date lack the capability to consider and address uncertainty, which demands that the robots be caged or confined to a dedicated, structured workspace. For example, success of thousands of mobile robots nowadays deployed in logistics centers is heavily reliant on their closed and controlled operating environments. In this thesis, we propose a series of computationally efficient algorithms that can collectively overcome uncertainty of various sources towards reliable autonomy for "cage-free" robotic operations. The methods presented in the thesis leverage probability theory to quantify the amount of present and future uncertainty. Based on the quantification, we develop planning and control algorithms that either mitigate, avoid the risk of, or are robust against uncertainty so that the robot can successfully accomplish a given task. We take a model-based approach in developing those algorithms, which allows us to exploit physical properties of dynamical systems and onboard sensors when possible. Another crucial aspect of the proposed methods is their online nature, meaning that control signals are computed in situ based on the currently available information. This is enabled by fast, efficient computation of our algorithms, and is advantageous in that the robot can quickly react to rapidly changing environments. In the first part of the thesis, we address challenges associated with state uncertainty, which represents unknowns about the current state of the system of interest. This can include unknown intent of other interacting agents, or positions of targets to locate. We propose and employ recursive Bayesian inference frameworks to keep track of evolving state uncertainty over time. The proposed planning algorithms further assist the inference frameworks to actively mitigate state uncertainty as appropriate, so that the robot can execute suitable control actions with certainty. We leverage tools from sequential decision-making and optimal control to develop those algorithms. We demonstrate the effectiveness of our approach in a multitude of tasks that involve state uncertainty, with different combinations of dynamical systems and sensing modalities. This includes vision-based active intent inference, active target tracking with range-only observations, and simultaneous object manipulation and parameter estimation. We then turn our attention to transition uncertainty, which governs the unpredictability of future states of the system. We especially focus on safety-critical problems where transition uncertainty must not be ignored. For instance, a robot navigating in close proximity to humans has to carefully perform planning so that collisions are avoided with high confidence. We take a risk-aware planning approach, in which a risk metric that takes into account the variance of uncertainty is to be optimized. While being computationally efficient, our proposed method does not require knowledge of the analytical form of the underlying probability distribution that quantifies transition uncertainty, nor is it limited to a certain class of distributions such as Gaussian. This atypical feature enables us to leverage modern data-driven generative models for uncertainty quantification. We demonstrate the applicability of our approach to the aforementioned robot navigation task, where we show that the proposed framework can safely navigate the robot towards its goal while interacting with more than 50 humans simultaneously in real time. Moreover, our risk-aware formulation is demonstrated to promote safety in both simulation and a real-world experiment, by inducing a proactive robot behavior that avoids risky situations where high variance of uncertainty could lead to imminent collision. The last part of this thesis considers model uncertainty, which is attributed to imperfect modeling of the underlying stochastic phenomena. Our approach makes the planner distributionally robust, in which the planner selects a control policy that acts against a worst-case distribution within an offline-computed set of plausible distributions that could quantify transition uncertainty. We develop a tractable algorithm leveraging mathematical equivalence between risk-aware planning and distributionally robust planning. We show in simulation that the proposed planning framework can safely avoid collision despite imperfect knowledge of the stochastic human motion model. Furthermore, our approach lets the risk-aware planner dynamically adjust the level of risk-sensitivity online, which further improves the flexibility of conventional risk-aware planning methods. The algorithms developed in this thesis will ultimately allow intelligent mobile robots to operate in considerably more uncertain and dynamic workspaces than the current industrial standard. This will open up possibilities for various practical applications, including autonomous field robots for persistent environmental monitoring, fully-automated driving on urban roads, and autonomous drone flights in densely populated areas for logistics services. We believe that such "cage-free" robotic operations will be enabled by proper consideration and treatment of uncertainty, and that our methods will pave the way towards more reliable robotic autonomy in open and interactive environments.

Planning algorithms are impacting technical disciplines and industries around the world, including robotics, computer-aided design, manufacturing, computer graphics, aerospace applications, drug design, and protein folding. Written for computer scientists and engineers with interests in artificial intelligence, robotics, or control theory, this is the only book on this topic that tightly integrates a vast body of literature from several fields into a coherent source for teaching and reference in a wide variety of applications. Difficult mathematical material is explained through hundreds of examples and illustrations.

A text that makes the mathematical underpinnings of robot motion accessible and relates low-level details of implementation to high-level algorithmic concepts. Robot motion planning has become a major focus of robotics. Research findings can be applied not only to robotics but to planning routes on circuit boards, directing digital actors in computer graphics, robot-assisted surgery and medicine, and in novel areas such as drug design and protein folding. This text reflects the great advances that have taken place in the last ten years, including sensor-based planning, probabalistic planning, localization and mapping, and motion planning for dynamic and nonholonomic systems. Its presentation makes the mathematical underpinnings of robot motion accessible to students of computer science and engineering, rleating low-level implementation details to high-level algorithmic concepts.

This work presents a behavior planning algorithm for automated driving in urban environments with an uncertain and dynamic nature. The algorithm allows to consider the prediction uncertainty (e.g. different intentions), perception uncertainty (e.g. occlusions) as well as the uncertain interactive behavior of the other agents explicitly. Simulating the most likely future scenarios allows to find an optimal policy online that enables non-conservative planning under uncertainty.

This book presents the recent research advances in linear and nonlinear control techniques. From both a theoretical and practical standpoint, motion planning and related control challenges are key parts of robotics. Indeed, the literature on the planning of geometric paths and the generation of time-based trajectories, while accounting for the compatibility of such paths and trajectories with the kinematic and dynamic constraints of a manipulator or a mobile vehicle, is extensive and rich in historical references. Path planning is vital and critical for many different types of robotics, including autonomous vehicles, multiple robots, and robot arms. In the case of multiple robot route planning, it is critical to produce a safe path that avoids colliding with objects or other robots. When designing a safe path for an aerial or underwater robot, the 3D environment must be considered. As the number of degrees of freedom on a robot arm increases, so does the difficulty of path planning. As a result, safe pathways for high-dimensional systems must be developed in a timely manner. Nonetheless, modern robotic applications, particularly those requiring one or more robots to operate in a dynamic environment (e.g., human–robot collaboration and physical interaction, surveillance, or exploration of unknown spaces with mobile agents, etc.), pose new and exciting challenges to researchers and practitioners. For instance, planning a robot's motion in a dynamic environment necessitates the real-time and online execution of difficult computational operations. The development of efficient solutions for such real-time computations, which could be offered by specially designed computational architectures, optimized algorithms, and other unique contributions, is thus a critical step in the advancement of present and future-oriented robotics.

Computer Science Workbench is a monograph series which will provide you with an in-depth working knowledge of current developments in computer technology. Every volume in this series will deal with a topic of importance in computer science and elaborate on how you yourself can build systems related to the main theme. You will be able to develop a variety of systems, including computer software tools, computer graphics, computer animation, database management systems, and computer-aided design and manufacturing systems. Computer Science Workbench represents an important new contribution in the field of practical computer technology. TOSIYASU L. KUNII To my parents Kenjiro and Nori Fujimura Preface Motion planning is an area in robotics that has received much attention recently. Much of the past research focuses on static environments - various methods have been developed and their characteristics have been well investigated. Although it is essential for autonomous intelligent robots to be able to navigate within dynamic worlds, the problem of motion planning in dynamic domains is relatively little understood compared with static problems.

An introduction to the techniques and algorithms of the newest field in robotics. Probabilistic robotics is a new and growing area in robotics, concerned with perception and control in the face of uncertainty. Building on the field of mathematical statistics, probabilistic robotics endows robots with a new level of robustness in real-world situations. This book introduces the reader to a wealth of techniques and algorithms in the field. All algorithms are based on a single overarching mathematical foundation. Each chapter provides example implementations in pseudo code, detailed mathematical derivations, discussions from a practitioner's perspective, and extensive lists of exercises and class projects. The book's Web site, www.probabilistic-robotics.org, has additional material. The book is relevant for anyone involved in robotic software development and scientific research. It will also be of interest to applied statisticians and engineers dealing with real-world sensor data.

The dominant theme of this book is to introduce the different path planning methods and present some of the most appropriate ones for robotic routing; methods that are capable of running on a variety of robots and are resistant to disturbances; being real-time, being autonomous, and the ability to identify high risk areas and risk management are the other features that will be mentioned in the introduction of the methods. The introduction of the profound significance of the robots and delineation of the navigation and routing theme is provided in the first chapter of the book. The second chapter is concerned with the subject of routing in unknown environments. In the first part of this chapter, the family of bug algorithms including are described. In the following, several conventional methods are submitted. The last part of this chapter is dedicated to the introduction of two recently developed routing methods. In Chapter 3, routing is reviewed in the known environment in which the robot either utilizes the created maps by extraneous sources or makes use of the sensor in order to prepare the maps from the local environment. The robot path planning relying on the robot vision sensors and applicable computing hardware are concentrated in the fourth chapter. The first part of this chapter deals with routing methods supported mapping capabilities. The second part manages the routing dependent on vision sensor typically known as the best sensor within the routing subject. The movement of two-dimensional robots with two or three degrees of freedom is analyzed within the third part of this chapter. In Chapter 5, the performance of a few of the foremost important routing methods initiating from the second to fourth chapters is conferred regarding the implementation in various environments. The first part of this chapter is engaged in the implementation of the algorithms Bug1, Bug2, and Distbug on the pioneering robot. In the second part, a theoretical technique is planned to boost the robot's performance in line with obstacle collision avoidance. This method, underlying the tangential escape, seeks to proceed the robot through various obstacles with curved corners. In the third and fourth parts of this chapter, path planning in different environments is preceded in the absence and the presence of danger space. Accordingly, four approaches, named artificial fuzzy potential field, linguistic technique, Markov decision making processes, and fuzzy Markov decision making have been proposed in two following parts and enforced on the Nao humanoid robot.

Ensuring the safety of autonomous vehicles during operation is a challenging task. Numerous factors such as process noise, sensor noise, incorrect model etc. can yield uncertainty in robot's state. Especially for tracked vehicles operating on rough terrain, vehicle slip due to vehicle terrain interaction affects the vehicle system significantly. In such cases, the motion planning of the autonomous vehicle must be performed robustly, considering the uncertain factors in advance of the real-time navigation. The primary contribution of this thesis is to present a robust optimal global planner for autonomous tracked vehicles operating in off-road terrain with uncertain slip. In order to achieve this goal, three tasks must be completed. First, the motion planner must be able to work efficiently under the non-holonomic vehicle system model. An approximate method is applied to the tracked vehicle system ensuring both optimality and efficiency. Second, the motion planner should ensure robustness. For this, a robust incremental sampling based motion planning algorithm (CC-RRT*) is combined with the LQG-MP algorithm. CC-RRT* yields the optimal and probabilistically feasible trajectory by using a chance constrained approach under the RRT* framework. LQG-MP provides the capability of considering the role of compensator in the motion planning phase and bounds the degree of uncertainty to appropriate size. Third, the effect of slip on the vehicle system must be modeled properly. This can be done in advance of operation if we have experimental data and full information about the environment. However, in case where such knowledge is not available, the online slip estimation can be performed using system identification method such as the IPEM algorithm. Simulation results shows that the resulting algorithms are efficient, optimal, and robust. The simulation was performed on a realistic scenario with several important factors that can increase the uncertainty of the vehicle. Experimental results are also provided to support the validity of the proposed algorithm. The proposed framework can be applied to other robotic systems where robustness is an important issue.

In recent years, robotic technology has improved significantly, aided by cutting-edge scientific research studies and innovative industrial designs. It has taken a progressive leap from the coordinated world of industry to the less-ordered domestic domain with great advancements in sensor technology and computational intelligence. It is beginning to prove more useful than a robot vacuum cleaner or a mere plaything in human-centric spaces. This has created an imminent need for robust intelligence for a robot to move optimally with high efficiency and collision-free navigation. This research provides valuable insights into all significant stages required for autonomous navigation in dynamic cluttered environments and makes several important contributions in the area.A unique and real-time method for global path planning and collision avoidance for navigation of a mobile robot in complex time varying environments is developed. An occupancy-based three dimensional (3D) grid map and model-based obstacle prediction are employed to represent the dynamic environment. Path planning and obstacle avoidance are performed by applying a cost-evaluation function on time-space Distance Transforms to uniquely produce the optimal path at the time of planning. Dealing with uncertainty with regard to the position of obstacles for a given navigation task is accommodated by introducing the notion of probabilities to the algorithm. The spatio-temporal cost evaluation based path planning algorithm provides the key contribution of this research.A robust method of pose estimation and tracking for a mobile robot is also investigated. The technique utilises an overhead panoramic vision camera in an indoor cluttered environment with the robot workspace of a two-dimensional planar surface. It is fast and does not require any unwarping of the panoramic view. A unique system, combining mean-shift, Kalman Filter and Hough Transform-based tracking, is used to improve the result. Experiments are conducted confirming that the system is capable of reliably localising and tracking the robot in cluttered scenes with variations of illumination and periods of occlusion.The thesis commences by describing the design of a real-time open-source 3D simulation platform based on a game engine. The platform is primarily aimed towards research in mobile robotics, in-game character manipulation, visual surveillance-related research and high quality synthetic video generation. It provided the initial test-bed for this research to analyse ideas and algorithms including path planning, prior to the physical realisation experiments.Finally, a complete navigation system is integrated for a wheel-based mobile robot verifying the innovations in a real-world scenario. The system will be incorporated into a larger project that is aimed towards the enhancement of robotic assistive technologies for elderly and disabled people.