Abstract: A system for avoiding obstacles in the path of a vehicle including a sensor assembly (24) having a field of view (575) with a plurality of sectors (576, 578, 580) for detecting the distance of objects within each sector. The system further includes an element (528) for identifying obstructed sectors in which objects are detected within a predetermined range and selects an unobstructed sector close in alignment to the direction of the path to designate around the object a clear path which is close to the original path.

Abstract: The Ground Surveillance Robot (GSR) project has proceeded continuously since the Fall of 1980, and in that time an autonomous vehicle design and some degree of implementation has been achieved. The vehicle design has been partitioned into sensor, control, and planning subsystems. A distributed blackboard scheme has been developed which provides the mechanism by which these subsystems are coordinated. Vehicle position and orientation are supplied by vehicle attitude and navigation sensor subsystems. Obstacle avoidance capability has been implemented by fusing information from vision and acoustic ranging sensors into local goals and avoidance points. The influence of these points is combined through potential field techniques to accomplish obstacle avoidance control. Distant terrain characteristics are identified using information from a gray-level vision system, a color vision system, and a computer-controlled laser ranging sensor. These characteristics are used by a general planning engine to develop the desired path to a visible goal in the direction of the final goal. Progress to the final goal consists of a succession of movements from one distant but visible intermediate goal to another. The experience from implementing this autonomous vehicle has indicated the need for an integrated set of debugging tools which make the faults in subsystem hardware and software more distinguishable.

Abstract: An obstacle data processing system for an unmanned self-controlled vehicle, capable of enabling the vehicle to automatically avert any obstacle which lies in the course of running of the vehicle. The system has obstacle position memory device capable of accumulating position data concerning specific obstacles and, hence, forming data concerning the distribution of the obstacles. The vehicle therefore can conduct appropriate averting operation in accordance with the distribution of a plurality of obstacles.

Abstract: Redundancy represents one key towards design and synthesis of more versatile manipulators. Obstacle avoidance and limited joint range constitute two kinds of constraints which can be potentially met by a kinematically redundant manipulator. The natural scenario is the inverse kinematic problem which is certainly a crucial point for robotic manipulator analysis and control. Based on a recently proposed dynamic solution technique, the inverse kinematic problem for redundant manipulators is solved in this paper. The kinematics of the manipulator is appropriately augmented in order to include the above mentioned constraints; the result is an efficient, fast dynamic algorithm which only makes use of the direct kinematics of the manipulator. Extensive simulation results illustrate the tracking performance for a given trajectory in the Cartesian space, while guaranteeing a collision-free trajectory and/or not violating a mechanical jointiimit.

Abstract: This report describes range data based obstacle avoidance techniques developed for use on an autonomous road-following robot vehicle. The purpose of these techniques is to detect and locate obstacles present in a road environment for navigation of a robot vehicle equipped with an active laser-based range sensor. Techniques are presented for obstacle detection, obstacle location, and coordinate transformations needed in the construction of Scene Models (symbolic structures representing the 3-D obstacle boundaries used by the vehicle's Navigator for path planning). These techniques have been successfully tested on an outdoor robotic vehicle, the Autonomous Land Vehicle (ALV), at speeds up to 3.5 km/hour.

Abstract: : We report on experiments with a mobile robot using one vision process (forward motion vision) to calibrate another (stereo vision) without resorting to any external units of measurement. Both are calibrated to a velocity dependent coordinate system which is natural to the task of obstacle avoidance. The foundations of these algorithms, in a world of perfect measurement, are quite elementary. The contribution of this work is to make them noise tolerant while remaining simple computationally. Both the algorithms and the calibration procedure are easy to implement and have shallow computational depth, making them (1) run at reasonable speed on moderate uni-processors, (2) appear practical to run continuously, maintaining an up-to-the-second calibration on a mobile robot, and (3) appear to be good candidates for massively parallel implementations.

Abstract: Resolved motion rate control of kinematically redundant manipulators using the simplest of the generalized inverses, the {1}-inverse, is investigated For the case of a planar articulated coordinate redundant robot with an arbitrary number of axes, closed-form equations for the joint rates are developed The structure of these equations is such that the redundant joint velocities can be arbitrarily specified Simple constraints are presented which ensure that closed trajectories in gripper configuration space generate closed trajectories in joint space A strategy for controlling the redundant joint rates that is applicable to obstacle avoidance is examined and illustrated with examples

Abstract: The primary vision task in road-following for a mobile robot is to provide a description of the road environment, including possible obstacles on the road. Techniques are presented for road segmentation and obstacle detection based on color video data. Using constraints on road characteristics in the image space and in 3D color space, the road is extracted and represented by its edges. Assuming vehicle movement, obstacles are detected at a distance and an obstacle avoidance mode is entered.

Abstract: An improved obstacle avoidance algorithm for the reflexive pilot in the FMC autonomous land vehicle is presented. To perform the obstacle avoidance operation, the specia] structure of the local map derived from sonic sensor information is used to find a collision-free path. This is done by first reducing the vehicle to a line model to obtain an equivalent problem with enlarged obstacles, and then dealing with the free space directly.

Abstract: The problem of real time robust control of robot manipulators in the presence of obstacles is considered. In order to achieve high speeds of operation in addition to obstacle avoidance, actuator dynamics and noise are modelled and added to the system description. The Obstacle Avoidance Strategy (OAS) translates each state constraint (obstacle) into state dependent control constraints (SDCC), which take the form of hyperplanes in the control space. The intersection of the sets defining the SDCC and the hard control bounds form a polygon in the control space. The Optimal Decision Stragety (ODS) is then used to calculate the control which lies in this polygon and minimizes the deviation between the acceleration vector of the CCR and a desired field of accelerations. Although the OAS algorithm is developed for a Cartesian Coordinate Robot (CCR), its approach can be utilized for articulated coordinate robots. Simulation results display the effectiveness of the algorithm for a workspace hosting multiple obstacles. The simplicity of the algorithm makes it desirable for real time control.

Abstract: This paper addresses an overview of time-varying obstacle avoidance problems for robotic manipulators. We propose four different approaches to solving these problems. They are the Heuristic Off-line (HOF) approach, the Heuristic On-line (HON) approach, the Analytic Off-line (AOF) approach, and the Analytic On-line (AON) approach. The AOF approach is particularly pursued and derived in this paper. Some fundamental difficulties are then discussed in the AOF approach. It is shown that the analytic approach is not always successful in solving the time-varying obstacle avoidance problems.

Abstract: This paper addresses an overview of time-varying obstacle avoidance problems for robotic manipulators. We propose four different approaches to solving these problems. They are the Heuristic Off-line (HOF) approach, the Heuristic On-line (HON) approach, the Analytic Off-line (AOF) approach, and the Analytic Online (AON) approach. The AOF approach is particularly pursued and derived in this paper. Some fundamental difficulties are then discussed in the AOF approach. It is shown that the analytic approach is not always successful in solving the time-varying obstacle avoidance problems.

Abstract: Findings are presented of a study to determine the feasibility of developing and demonstrating a long range autonomous underwater vehicle. Based on a real world scale program need, a technology development and capability demonstration program is described. The program objectives necessary to provide a proof of principle including expected system performance capabilities are described together with an activity program for the demonstration system. Sensor systems for navigation, obstacle avoidance, passive detection, vehicle motion and vehicle health are described. Particular attention is paid to the discussion of the hardware and software architecture for the system with an emphasis on providing as much top-down guidance as possible and to exploit sensor modality differences to produce complementary perceptual processes in the system. The discussion of the software includes the application of a system capable of supporting parallelism in its knowledge source modules and a organized collection of perceptual and navigation modules tied together through a blackboard. The paper describes the database/communication system, the AUV and system block diagram together with the issues which are inherent in the integration of the multiple sensors of the system. Path planning abilities are described against a background of actual sonar-depth data obtained during the study. Simulations of a proposed vehicle, including six degrees of freedom, in a marine environment are described. The evolution of the AUV system from simulation through component testing to the at sea demonstration is discussed.

Abstract: This paper addresses the critical issues of sense acquisition and sense analysis, using multiple Obstacle Avoidance (OA) sensors, for Autonomous Underwater Vehicles (AUVs). Currently, an AUV research and development testbed is being engineered at Martin Marietta Baltimore Aerospace, and the pertinent research on OA sensing that may be applied to such a testbed is presented. The complexities of relevant and conventional navigation systems for undersea vehicles are also discussed. Using a multitude of sensors and the Zonal-Spot Environmental Analysis (ZSEA) sensing technique, the spatial scenarios are characterized cogently by amalgamating the sensor information to form a description of the external world, which in turn are preserved in a world model database. The ZSEA sensing strategy performs sensor-level processes to recognize obstacles with certain levels of surety, based on apriori semantic data, and thereby provides safer and more efficient path planning capabilities to the AUV. Furthermore, the ZSEA sensing strategy ameliorates sensory deprivation and paves the way for the incremental enhancement of the control process at the higher levels without having to modify the lower servocontrol levels.

Abstract: 下等動物の行動学をヒントにした自律移動ロボットの一戦略SbS (サインパターン定形行動連鎖) に基づいて, 晴信-4号機を開発中である.晴信-4号機は頭部と胸部に2台のカラーTVカメラを有し, 68020VMEシステムを塔載している. (1) 構内道路の移動を2つの定形行動, 即ち, 道路境界や花壇等のサインパターンにする沿目標移動と柱や門, 街路樹等をサインパターンにする向目標移動, で実現する. (2) 出発点から目標地点までの移動はサインパターンと定形行動のリストからなる走行指令書で定義する. (3) 障害物回避や段差乗越しも定形行動で行なう. (4) 超音波センサやバンパーセンサによる衝突防止も定形行動で行なう. (5) 路上の影と障害物の区別は上下2台のTVカメラのステレオ視で行なう.一つの道路シーンの画像処理所要時間は障害物検出を含めて2秒以下である.

Abstract: Dextrous hand designs require planning and coordination of multijointed fingers. This communication presents a dynamic programming algorithm to optimize the trajectory of a finger or a fingerlike manipulator. The algorithm generates a minimum cost trajectory for a specified cost function (such as accuracy or time of travel) and a set of constraints (such as obstacles along the trajectory). We present simulations demonstrating the trajectory planning for simplified fingerlike manipulator configurations. This approach is potentially applicable to several additional situations faced by multijointed fingers employed in dextrous hand designs, such as complex trajectory planning, obstacle avoidance, and fault tolerance. We illustrate some results obtained by computer simulations of these problems.

Abstract: Euclidean distance is the most popular criterion for robot path planning. However, the shortest path (SP) is dangerous in some cases because such a path drives the robot too close to obstacles. When safety is the main concern, a center-line path (CLP) providing the maximum clearance from obstacles would be preferable over the SP, although the length of a CLP could be considerably longer than that of a SP. Since the SP and CLP are two extremes with respect to the distance and safety criteria, respectively, it would be useful in practice to strike a compromise between the two criteria. The purpose of this paper is to develop a method for determining an optimal path with a weighted distance-safety criterion. The method is composed of three parts: (i) construction of a region map by dividing the workspace, (ii) inter-region optimization to determine the entry and departure points of the path in each region, and (iii) intra-region optimization for determining the (optimal) path segment within each region. The region map is generated by using an approximate Voronoi diagram, and the inter- (intra-) region optimization is achieved by using the variational dynamic programming. Although it is developed for 2D problems, our method can be easily extended to a class of 3D problems. Numerical examples are also presented to demonstrate the method.

Abstract: Range imagery from a laser scanner can be used to provide sufficient information for docking and obstacle avoidance procedures to be performed automatically. Three dimensional model-based computer vision algorithms in development can perform these tasks even with targets which may not be cooperative (that is, objects without special targets or markers to provide unambiguous location points). Roll, pitch and yaw of the vehicle can be taken into account as image scanning takes place, so that these can be corrected when the image is converted from egocentric to world coordinates. Other attributes of the sensor, such as the registered reflectence and texture channels, provide additional data sources for algorithm robustness. Temporal fusion of sensor immages can take place in the work coordinate domain, allowing for the building of complex maps in three dimensional space.

Abstract: : ROBART II is a battery powered autonomous sentry robot used by the Naval Ocean Systems Center in San Diego as a testbed in the research areas of environmental modeling and intelligent security assessment. An architecture of nine distributed microprocessors makes possible advanced control strategies and real-time data acquisition capability. Higher level functionality (map generation, path planning, position estimation, obstacle avoidance and statistical security assessment) is currently implemented by the Planner on an IBM-PC/AT computer, using a radio link for communication with the Scheduler. Numerous sensors are incorporated into the system to yield appropriate information for use in collision avoidance, navigational planning, environmental awareness, assessing terrain traversability, and performing security related functions. Two separate drive motors provide for differential steering, allowing the robot to turn in place in order to maneuver in congested indoor environments. The entire unit is housed in a rugged plastic and fiberglass body, measuring 17 inches wide and 23 inches long at the base, and extending to a height of 50 inches. Special internal circuitry checkpoints are analyzed by self-diagnostic software, and operator assistance is requested if necessary through speech synthesis. Keywords: Robotics; Artificial intelligence.

Abstract: Mobile mechanical devices working with inspection and maintenance of pipes in electrical power plants might become useful to E.D.F. A specific field in mobile robotics, with its proper know-how and complications, is being developed. Simulations of different aspects of robot conception are presented : - environment modelling, - kinetics : modelling which can include closed chains, - navigation and obstacle avoidance : sensor choice and positionning, strategy control.

Abstract: This paper describes the functionality of a local planner module for autonomous mobile robots. As an intermediate layer of a hierarchical planning system for autonomous vehicles, this module translates goals provided by a map-based planner into general vehicle action modes called activities. Activities are composed of pre-defined sets of reflexive behaviors which yield known performance characteristics under certain specific environmental conditions. Activity selection is performed on the basis of perception-supplied descriptions of the local environment, descriptions of expected terrain characteristics from a map-based planner, local map information accumulated by the local planner itself, and knowledge stored with each activity describing how that activity may achieve various goals and handle various failures. Local planner reasoning occurs within a monitoring process associated with the selected activity. Selected activities communicate with the local planner by posting messages onto a blackboard. These messages convey information to the local planner such as the distance to the nearest obstacle in a certain direction and activity completion status. As the vehicle moves, the local planner builds a map that records the path traversed, as well as major features and landmarks encountered. This information is used when the vehicle needs to backtrack. The local planner is implemented in Lisp, and has been demonstrated in a simulated environment.

Abstract: This paper focuses on concepts and technologies required to develop a robotic air vehicle (RAV). A vehicle of this type has the capability to be a launch and forget weapon system. The authors are engineers and pilots so they view both the technical approach and piloting issues with equal importance. RAV must have the machine intelligence to make decisions within the mission and battlefield constraints. This requires a piloting expert system and route planner to perform passive terrain following, terrain avoidance, obstacle avoidance, and autonomous navigation based on low cost sensor inputs such as a multifunction FLIR, digital terrain map, and directional reference systems. RAV is a cost effective way to fight in a threat environment where aircrew loss rates would be unacceptable. RAV provides the Air Force a means to expand its combat capabilities.

Abstract: The navigation problem for autonomous land vehicles is formulated mathematically using optimal control theory so that obstacle avoidance, terrain negotiation and goal attainment can be treated within one general framework. A simple example is used to illustrate this approach.

Abstract: This paper presents various projects which are being developed at the Institute for Informatics at the Robotics Research Group. Several of these projects will be integrated to support the development of an autonomous assembly robot. With its two arms under sensor guidance, the robot will be capable of assembling a standard benchmark (workpieces). To achieve mobility, the robot will be equipped with a self-propelled platform. Several vision systems are employed in order to support the navigation of the robot and the assembly task. An AI-based system is being developed for assembly planning, implicit programming, route planning (obstacle avoidance), assembly monitoring, and error recovery.

Abstract: This paper describes a 3-dimensional vision system that utilizes two views of a scene and textural information in determining distances to objects in a scene. The application of the system lies with determining distances to obstacles in mobile robot guidance. The analysis of a scene is performed in two stages. First, orientations of planar textured surfaces in a scene are estimated using textural information in a single image frame. This step determines the relative positions of obstacles located on textured surfaces. In the second step, exact positions of the obstacles are determined using two views of a scene.

Abstract: Typically a robot cannot handle a situation in which it has to avoid a random obstacle in its path. This paper discusses an approach for on-line obstacle avoidance using a real-time expert system. The method relies on an algorithm which employs a one-step optimization with N-step lookahead and global linearization. The system receives input data from a vision system, which is used to identify the obstacles and their shape. An expert system observes the movements of the robot and takes over the control when there is a possibility of collision. The system employs the concept of focusing, moving faster when there is space, and slowing down to give more time to solve the problem of bypassing when operating space is restricted.

Abstract: The Space Station Program marks a new era in space exploration and habitation. To meet the challenges of this new era, more extensive use of remote manipulation and robotics is expected. This paper describes a new space telerobot concept which addresses both teleoperations and robotics needs of future space programs, while merging the desirable characteristics of both technologies. This new concept is based on knowledge and experience gained from manipulator systems developed to meet the needs of remote nuclear applications. It merges desirable characteristics of teleoperation and robotic technologies. Presented here are design goals for the telerobot, a description of the mechanical and control abilities, and applications for earth and space. The concept incorporates mechanical traction drives, redundant kinematics, and modular arm subelements to provide a backlash-free manipulator capable of obstacle avoidance. Further development of this telerobot is in progress at the Oak Ridge National Laboratory.