Jean-Claude Latombe

Probabilistic roadmaps for path planning in high-dimensional configuration spaces

A new motion planning method for robots in static workspaces is presented. This method proceeds in two phases: a learning phase and a query phase. In the learning phase, a probabilistic roadmap is constructed and stored as a graph whose nodes correspond to collision-free configurations and whose edges correspond to feasible paths between these configurations. These paths are computed using a simple and fast local planner. In the query phase, any given start and goal configurations of the robot are connected to two nodes of the roadmap; the roadmap is then searched for a path joining these two nodes. The method is general and easy to implement. It can be applied to virtually any type of holonomic robot. It requires selecting certain parameters (e.g., the duration of the learning phase) whose values depend on the scene, that is the robot and its workspace. But these values turn out to be relatively easy to choose, Increased efficiency can also be achieved by tailoring some components of the method (e.g., the local planner) to the considered robots. In this paper the method is applied to planar articulated robots with many degrees of freedom. Experimental results show that path planning can be done in a fraction of a second on a contemporary workstation (/spl ap/150 MIPS), after learning for relatively short periods of time (a few dozen seconds).

Randomized Kinodynamic Motion Planning with Moving Obstacles

This paper presents a novel randomized motion planner for robots that must achieve a specified goal under kinematic and/or dynamic motion constraints while avoiding collision with moving obstacles with known trajectories. The planner encodes the motion constraints on the robot with a control system and samples the robots state × time space by picking control inputs at random and integrating its equations of motion. The result is a probabilistic roadmap of sampled state ×time points, called milestones, connected by short admissible trajectories. The planner does not precompute the roadmap; instead, for each planning query, it generates a new roadmap to connect an initial and a goal state×time point. The paper presents a detailed analysis of the planners convergence rate. It shows that, if the state×time space satisfies a geometric property called expansiveness, then a slightly idealized version of our implemented planner is guaranteed to find a trajectory when one exists, with probability quickly converging to 1, as the number of milestones increases. Our planner was tested extensively not only in simulated environments, but also on a real robot. In the latter case, a vision module estimates obstacle motions just before planning starts. The planner is then allocated a small, fixed amount of time to compute a trajectory. If a change in the expected motion of the obstacles is detected while the robot executes the planned trajectory, the planner recomputes a trajectory on the fly. Experiments on the real robot led to several extensions of the planner in order to deal with time delays and uncertainties that are inherent to an integrated robotic system interacting with the physical world.

Randomized preprocessing of configuration for fast path planning

This paper presents a new approach to path planning for robots with many degrees of freedom (DOF) operating in known static environments. The approach consists of a preprocessing and a planning stage. Preprocessing, which is done only once for a given environment, generates a network of randomly, but properly selected, collision-free configurations (nodes). Planning then connects any given initial and final configurations of the robot to two nodes of the network and computes a path through the network between these two nodes. Experiments show that after paying the preprocessing cost (on the order of hundreds of seconds), planning is extremely fast (on the order of a fraction of a second for many difficult examples involving a 10-DOF robot). The approach is particularly attractive for many-DOF robots which have to perform many successive point-to-point motions in the same environment.<<ETX>>

Analysis of probabilistic roadmaps for path planning

Provides an analysis of a path planning method which uses probabilistic roadmaps. This method has proven very successful in practice, but the theoretical understanding of its performance is still limited. Assuming that a path /spl gamma/ exists between two configurations a and b of the robot, we study the dependence of the failure probability to connect a and b on (i) the length of /spl gamma/, (ii) the distance function of /spl gamma/ from the obstacles, and (iii) the number of nodes N of the probabilistic roadmap constructed. Importantly, our results do not depend strongly on local irregularities of the configuration space, as was the case with previous analysis. These results are illustrated with a simple but illuminating example. In this example, we provide estimates for N, the principal parameter of the method, in order to achieve failure probability within prescribed bounds. We also compare, through this example, the different approaches to the analysis of the planning method.

Path planning in expansive configuration spaces

We introduce the notion of expansiveness to characterize a family of robot configuration spaces whose connectivity can be effectively captured by a roadmap of randomly-sampled milestones. The analysis of expansive configuration spaces has inspired us to develop a new randomized planning algorithm. This algorithm tries to sample only the portion of the configuration space that is relevant to the current query, avoiding the cost of precomputing a roadmap for the entire configuration space. Thus, it is well-suited for problems where a single query is submitted for a given environment. The algorithm has been implemented and successfully applied to complex assembly maintainability problems from the automotive industry.

Planning motions with intentions

We apply manipulation planning to computer animation. A new path planner is presented that automatically computes the collision-free trajectories for several cooperating arms to manipulate a movable object between two configurations. This implemented planner is capable of dealing with complicated tasks where regrasping is involved. In addition, we present a new inverse kinematics algorithm for the human arms. This algorithm is utilized by the planner for the generation of realistic human arm motions as they manipulate objects. We view our system as a tool for facilitating the production of animation.

Navigation Strategies for Exploring Indoor Environments

In this paper, we investigate safe and efficient map-building strategies for a mobile robot with imperfect control and sensing. In the implementation, a robot equipped with a range sensor builds apolygonal map (layout) of a previously unknown indoor environment. The robot explores the environment and builds the map concurrently by patching together the local models acquired by the sensor into a global map. A well-studied and related problem is the simultaneous localization and mapping (SLAM) problem, where the goal is to integrate the information collected during navigation into the most accurate map possible. However, SLAM does not address the sensor-placement portion of the map-building task. That is, given the map built so far, where should the robot go next? This is the main question addressed in this paper. Concretely, an algorithm is proposed to guide the robot through a series of “good” positions, where “good” refers to the expected amount and quality of the information that will be revealed at each new location. This is similar to the next-best-view (NBV) problem studied in computer vision and graphics. However, in mobile robotics the problem is complicated by several issues, two of which are particularly crucial. One is to achieve safe navigation despite an incomplete knowledge of the environment and sensor limitations (e.g., in range and incidence). The other issue is the need to ensure sufficient overlap between each new local model and the current map, in order to allow registration of successive views under positioning uncertainties inherent to mobile robots. To address both issues in a coherent framework, in this paper we introduce the concept of a safe region, defined as the largest region that is guaranteed to be free of obstacles given the sensor readings made so far. The construction of a safe region takes sensor limitations into account. In this paper we also describe an NBV algorithm that uses the safe-region concept to select the next robot position at each step. The new position is chosen within the safe region in order to maximize the expected gain of information under the constraint that the local model at this new position must have a minimal overlap with the current global map. In the future, NBV and SLAM algorithms should reinforce each other. While a SLAM algorithm builds a map by making the best use of the available sensory data, an NBV algorithm, such as that proposed here, guides the navigation of the robot through positions selected to provide the best sensory inputs.

Nonholonomic multibody mobile robots: Controllability and motion planning in the presence of obstacles

We consider mobile robots made of a single body (car-like robots) or several bodies (tractors towing several trailers sequentially hooked). These robots are known to be nonholonomic, i.e., they are subject to nonintegrable equality kinematic constraints involving the velocity. In other words, the number of controls (dimension of the admissible velocity space), is smaller than the dimension of the configuration space. In addition, the range of possible controls is usually further constrained by inequality constraints due to mechanical stops in the steering mechanism of the tractor. We first analyze the controllability of such nonholonomic multibody robots. We show that the well-known Controllability Rank Condition Theorem is applicable to these robots even when there are inequality constraints on the velocity, in addition to the equality constraints. This allows us to subsume and generalize several controllability results recently published in the Robotics literature concerning nonholonomic mobile robots, and to infer several new important results. We then describe an implemented planner inspired by these results. We give experimental results obtained with this planner that illustrate the theoretical results previously developed.

A Single-Query Bi-Directional Probabilistic Roadmap Planner with Lazy Collision Checking

This paper describes a new probabilistic roadmap (PRM) path planner that is: (1) single-query — instead of pre-computing a roadmap covering the entire free space, it uses the two input query configurations as seeds to explore as little space as possible; (2) bidirectional — it explores the robot’s free space by concurrently building a roadmap made of two trees rooted at the query configurations; (3) adaptive — it makes longer steps in opened areas of the free space and shorter steps in cluttered areas; and (4) lazy in checking collision — it delays collision tests along the edges of the roadmap until they are absolutely needed. Experimental results show that this combination of techniques drastically reduces planning times, making it possible to handle difficult problems, including multi-robot problems in geometrically complex environments.

Motion Planning: A Journey of Robots, Molecules, Digital Actors, and Other Artifacts

During the past three decades, motion planning has emerged as a crucial and productive research area in robotics. In the mid-1980s, the most advanced planners were barely able to compute collision-free paths for objects crawling in planar workspaces. Today, planners efficiently deal with robots with many degrees of freedom in complex environments. Techniques also exist to generate quasioptimal trajectories, coordinate multiple robots, deal with dynamic and kinematic constraints, and handle dynamic environments. This paper describes some of these achievements, presents new problems that have recently emerged, discusses applications likely to motivate future research, and finally gives expectations for the coming years. It stresses the fact that nonrobotics applications (e.g., graphic animation, surgical planning, computational biology) are growing in importance and are likely to shape future motion-planning research more than robotics itself.