In recent years, autonomous driving has become an increasingly practical technology. With state-of-the-art computer and sensor engineering, autonomous vehicles may be produced and widely used for travel and logistics in the near future. They have great potential to reduce traffic accidents, improve transportation efficiency, and release people from driving tasks while commuting. Researchers have built autonomous vehicles that can drive on public roads and handle normal surrounding traffic and obstacles. However, in situations like lane changing and merging, the autonomous vehicle faces the challenge of performing smooth interaction with human-driven vehicles. To do this, autonomous vehicle intelligence still needs to be improved so that it can better understand and react to other human drivers on the road. In this thesis, we argue for the importance of implementing ”socially cooperative driving”, which is an integral part of everyday human driving, in autonomous vehicles. An intention-integrated Prediction- and Cost function-Based algorithm (iPCB) framework is proposed to enable an autonomous vehicles to perform cooperative social behaviors. We also propose a behavioral planning framework to enable the socially cooperative behaviors with the iPCB algorithm. The new architecture is implemented in an autonomous vehicle and can coordinate the existing Adaptive Cruise Control (ACC) and Lane Centering interface to perform socially cooperative behaviors. The algorithm has been tested in over 500 entrance ramp and lane change scenarios on public roads in multiple cities in the US and over 10; 000 in simulated case and statistical testing. Results show that the proposed algorithm and framework for autonomous vehicle improves the performance of autonomous lane change and entrance ramp handling. Compared with rule-based algorithms that were previously developed on an autonomous vehicle for these scenarios, over 95% of potentially unsafe situations are avoided.

La robotique humanoïde arrive a maturité avec des robots plus rapides et plus précis. Pour faire face à la complexité mécanique, la recherche a commencé à regarder au-delà du cadre habituel de la robotique, vers les sciences de la vie, afin de mieux organiser le contrôle du mouvement. Cette thèse explore le lien entre mouvement humain et le contrôle des systèmes anthropomorphes tels que les robots humanoïdes. Tout d’abord, en utilisant des méthodes classiques de la robotique, telles que l’optimisation, nous étudions les principes qui sont à la base de mouvements répétitifs humains, tels que ceux effectués lorsqu’on joue au yoyo. Nous nous concentrons ensuite sur la locomotion en nous inspirant de résultats en neurosciences qui mettent en évidence le rôle de la tête dans la marche humaine. En développant une interface permettant à un utilisateur de commander la tête du robot, nous proposons une méthode de contrôle du mouvement corps-complet d’un robot humanoïde, incluant la production de pas et permettant au corps de suivre le mouvement de la tête. Cette idée est poursuivie dans l’étude finale dans laquelle nous analysons la locomotion de sujets humains, dirigée vers une cible, afin d’extraire des caractéristiques du mouvement sous forme invariants. En faisant le lien entre la notion “d’invariant” en neurosciences et celle de “tâche cinématique” en robotique humanoïde, nous développons une méthode pour produire une locomotion réaliste pour d’autres systèmes anthropomorphes. Dans ce cas, les résultats sont illustrés sur le robot humanoïde HRP2 du LAAS-CNRS. La contribution générale de cette thèse est de montrer que, bien que la planification de mouvement pour les robots humanoïdes peut être traitée par des méthodes classiques de robotique, la production de mouvements réalistes nécessite de combiner ces méthodes à l’observation systématique et formelle du comportement humain. / Humanoid robotics is coming of age with faster and more agile robots. To compliment the physical complexity of humanoid robots, the robotics algorithms being developed to derive their motion have also become progressively complex. The work in this thesis spans across two research fields, human neuroscience and humanoid robotics, and brings some ideas from the former to aid the latter. By exploring the anthropological link between the structure of a human and that of a humanoid robot we aim to guide conventional robotics methods like local optimization and task-based inverse kinematics towards more realistic human-like solutions. First, we look at dynamic manipulation of human hand trajectories while playing with a yoyo. By recording human yoyo playing, we identify the control scheme used as well as a detailed dynamic model of the hand-yoyo system. Using optimization this model is then used to implement stable yoyo-playing within the kinematic and dynamic limits of the humanoid HRP-2. The thesis then extends its focus to human and humanoid locomotion. We take inspiration from human neuroscience research on the role of the head in human walking and implement a humanoid robotics analogy to this. By allowing a user to steer the head of a humanoid, we develop a control method to generate deliberative whole-body humanoid motion including stepping, purely as a consequence of the head movement. This idea of understanding locomotion as a consequence of reaching a goal is extended in the final study where we look at human motion in more detail. Here, we aim to draw to a link between “invariants” in neuroscience and “kinematic tasks” in humanoid robotics. We record and extract stereotypical characteristics of human movements during a walking and grasping task. These results are then normalized and generalized such that they can be regenerated for other anthropomorphic figures with different kinematic limits than that of humans. The final experiments show a generalized stack of tasks that can generate realistic walking and grasping motion for the humanoid HRP-2. The general contribution of this thesis is in showing that while motion planning for humanoid robots can be tackled by classical methods of robotics, the production of realistic movements necessitate the combination of these methods with the systematic and formal observation of human behavior.

In this thesis improvements, implementation and evaluation have been done on a motion planning algorithm for a full-sized reversing truck and trailer system. The motion planner is based on a motion planning algorithm called Closed-Loop Rapidly-exploring Random Tree (CL-RRT). An important property for a certain class of systems, stating that by selecting the input signals in a certain way the same result as reversing the time can be archived, is also presented. For motion planning this means that the problem of reversing from position A to position B can also be solved by driving forward from B to A and then reverse the solution. The use of this result in the motion planner has been evaluated and has shown to be very useful. The main improvements made on the CL-RRT algorithm are a faster collision detection method, a more efficient way to draw samples and a more correct heuristic cost-to-go function. A post optimizing or smoothing method that brings the system to the exact desired configuration, based on numerical optimal control, has also been developed and implemented with successful results. The motion planner has been implemented and evaluated on a full-scale truck with a dolly steered trailer prepared for autonomous operation with promising results.

Trajectory planning is an important component of autonomous driving. It takes the result of route-level navigation plan and generates the motion-level commands that steer an autonomous passenger vehicle (APV). Prior work on solving this problem uses either a sampling-based or optimization-based trajectory planner, accompanied by some high-level rule generation components.

L'objet de cette thèse est le développement et l'étude des algorithmes de planification de mouvement pour les systèmes hautement dimensionnés que sont les robots humanoïdes et les acteurs virtuels. Plusieurs adaptations des méthodes génériques de planification de mouvement randomisées sont proposées et discutées. Une première contribution concerne l'utilisation de techniques de réduction de dimension linéaire pour accélérer les algorithmes d'échantillonnage. Cette méthode permet d'identifier en ligne quand un processus de planification passe par un passage étroit de l'espace des configurations et adapte l'exploration en fonction. Cet algorithme convient particulièrement bien aux problèmes difficiles de la planification de mouvement pour l'animation graphique. La deuxième contribution est le développement d'algorithmes randomisés de planification sous contraintes. Il s'agit d'une intégration d'outils de cinématique inverse hiérarchisée aux algorithmes de planification de mouvement randomisés. On illustre cette méthode sur différents problèmes de manipulation pour robots humanoïdes. Cette contribution est généralisée à la planification de mouvements corps-complet nécessitant de la marche. La dernière contribution présentée dans cette thèse est l'utilisation des méthodes précédentes pour résoudre des tâches de manipulation complexes par un robot humanoïde. Nous présentons en particulier un formalisme destiné à représenter les informations propres à l'objet manipulé utilisables par un planificateur de mouvement. Ce formalisme est présenté sous le nom d'« objets documentés». / This thesis deals with the development and analysis of motion planning algorithms for high dimensional systems: humanoid robots and digital actors. Several adaptations of generic randomized motion planning methods are proposed and discussed. A first contribution concerns the use of linear dimensionality reduction techniques to speed up sampling algorithms. This method identifies on line when a planning process goes through a narrow passage of some configuration space, and adapts the exploration accordingly. This algorithm is particularly suited to difficult problems of motion planning for computer animation. The second contribution is the development of randomized algorithms for motion planning under constraints. It consists in the integration of prioritized inverse kinematics tools within randomized motion planning. We demonstrate the use of this method on different manipulation planning problems for humanoid robots. This contribution is generalized to whole-body motion planning with locomotion. The last contribution of this thesis is the use of previous methods to solve complex manipulation tasks by humanoid robots. More specifically, we present a formalism that represents information specific to a manipulated object usable by a motion planner. This formalism is presented under the name of "documented object".

"We address the problem of performing complex tasks for a robot operating in changing environments. We propose two approaches to the following problem: 1) define task-specific cost functions for motion planning that represent path quality by learning from an expert's preferences and 2) using constraint-based representation of the task inside learning from demonstration paradigm. In the first approach, we generate a set of paths for a given task using a motion planner and collect data about their features (path length, distance from obstacles, etc.). We provide these paths to an expert as a set of pairwise comparisons. We then form a ranking of the paths from the expert's comparisons. This ranking is used as training data for learning algorithms, which attempt to produce a cost function that maps path feature values to a cost that is consistent with the expert's ranking. We test our method on two simulated car-maintenance tasks with the PR2 robot: removing a tire and extracting an oil filter. We found that learning methods which produce non-linear combinations of the features are better able to capture expert preferences for the tasks than methods which produce linear combinations. This result suggests that the linear combinations used in previous work on this topic may be too simple to capture the preferences of experts for complex tasks. In the second approach, we propose to introduce a constraint-based description of the task that can be used together with the motion planner to produce the trajectories. The description is automatically created from the demonstration by performing segmentation and extracting constraints from the motion. The constraints are represented with the Task Space Regions (TSR) that are extracted from the demonstration and used to produce a desired motion. To account for the parts of the motion where constraints are different a segmentation of the demonstrated motion is performed using TSRs. The proposed approach allows performing tasks on robot from human demonstration in changing environments, where obstacle distribution or poses of the objects could change between demonstration and execution. The experimental evaluation on two example motions was performed to estimate the ability of our approach to produce the desired motion and recover a demonstrated trajectory."

This thesis addresses the problem of generating smooth and efficiently executable locomotion trajectories for legged robots under contact constraints. In addition, we want the trajectories to have the property that small changes in the foot position generate small changes in the joint target path. The first part of this thesis explores methods to select poses for a legged robot that maximises the workspace reachability while maintaining stability and contact constraints. It also explores methods to select configurations based on a reduced-dimensional search of the configuration space. The second part analyses time scaling strategy which tries to minimize the execution time while obeying the velocity and acceleration constraints. These two parts effectively result in smooth feasible trajectories for legged robots. Experiments on the RoboSimian robot demonstrate the effectiveness and scalability of the strategies described for walking and climbing on a rock climbing wall.

The daVinci Research Kit (dVRK) is a research platform that consists of the clinical daVinci surgical robot, provided by Intuitive Surgical to Academic Institutions. It provides an open source software and hardware platform for researchers to study and analyze the current architecture and expand the capabilities of the existing technology. The line between general purpose robotics and medical robotics has segregated the two fields. A significant part of the segregation lies at the software end, where new tools and methods developed in general purpose robotics cannot make it to medical robotics in a short amount of time. This research focuses on the integration of a widely used software architecture for general purpose robotics with the dVRK with the hope of utilizing the research and development from one field to the other. As a first step towards this bridging, a motion planning framework and a dynamic simulator has been developed for the dVRK using ROS. The motion planning framework is aimed to assist the surgeon in performing task with additional safety and machine intelligence. A few use cases have been proposed as well. Lastly, a Matlab Interface has been developed that is standalone in terms of usage and provides capabilities to interact with dVRK.

In needle-based surgical interventions, accurate alignment and insertion of the tool is paramount for providing proper treatment at a target site while minimizing healthy tissue damage. While manually-aligned interventions are well-established, robotics platforms promise to reduce procedure time, increase precision, and improve patient comfort and survival rates. Conducting interventions in an MRI scanner can provide real-time, closed-loop feedback for a robotics platform, improving its accuracy, yet the tight environment potentially impairs motion, and perceiving this limitation when planning a procedure can be challenging. This project developed a surgical workflow and software system for evaluating the workspace and planning the motions of a robotics platform within the confines of an MRI scanner. 3D Slicer, a medical imaging visualization and processing platform, provided a familiar and intuitive interface for operators to quickly plan procedures with the robotics platform over OpenIGTLink. Robotics tools such as ROS and MoveIt! were utilized to analyze the workspace of the robot within the patient and formulate the motion planning solution for positioning of the robot during surgical procedures. For this study, a 7 DOF robot arm designed for ultrasonic ablation of brain tumors was the targeted platform. The realized system successfully yielded prototype capabilities on the neurobot for conducting workspace analysis and motion planning, integrated systems using OpenIGTLink, provided an opportunity to evaluate current software packages, and informed future work towards production-grade medical software for MRI-guided, needle-based robotic interventions.

"When humans interact, there are many avenues of physical communication available ranging from vocal to physical gestures. In our past observations, when humans collaborate on manipulation tasks in shared workspaces there is often minimal to no verbal or physical communication, yet the collaboration is still fluid with minimal interferences between partners. However, when humans perform similar tasks in the presence of a robot collaborator, manipulation can be clumsy, disconnected, or simply not human-like. The focus of this work is to leverage our observations of human-human interaction in a robot's motion planner in order to facilitate more safe, efficient, and human-like collaborative manipulation in shared workspaces. We first present an approach to formulating the cost function for a motion planner intended for human-robot collaboration such that robot motions are both safe and efficient. To achieve this, we propose two factors to consider in the cost function for the robot's motion planner: (1) Avoidance of the workspace previously-occupied by the human, so robot motion is safe as possible, and (2) Consistency of the robot's motion, so that the motion is predictable as possible for the human and they can perform their task without focusing undue attention on the robot. Our experiments in simulation and a human-robot workspace sharing study compare a cost function that uses only the first factor and a combined cost that uses both factors vs. a baseline method that is perfectly consistent but does not account for the human's previous motion. We find using either cost function we outperform the baseline method in terms of task success rate without degrading the task completion time. The best task success rate is achieved with the cost function that includes both the avoidance and consistency terms. Next, we present an approach to human-attention aware robot motion generation which attempts to convey intent of the robot's task to its collaborator. We capture human attention through the combined use of a wearable eye-tracker and motion capture system. Since human attention isn't static, we present a method of generating a motion policy that can be queried online. Finally, we show preliminary tests of this method."