Showing papers on "Obstacle published in 1991"

Stepping over obstacles: gait patterns of healthy young and old adults

Abstract: Falls associated with tripping over an obstacle can be devastating to elderly individuals, yet little is known about the strategies used for stepping over obstacles by either old or young adults. The gait of gender-matched groups of 24 young and 24 old healthy adults (mean ages 22 and 71 years) was studied during a 4 m approach to and while stepping over obstacles of 0, 25, 51, or 152 mm height and in level obstacle-free walking. Optoelectronic cameras and recorders were used to record approach and obstacle crossing speeds as well as bilateral lower extremity kinematic parameters that described foot placement and movement trajectories relative to the obstacle. The results showed that age had no effect on minimum swing foot clearance (FC) over an obstacle. For the 25 mm obstacle, mean FC was 64 mm, or approximately three times that used in level gait; FC increased nonlinearly with obstacle height for all subjects. Although no age differences were found in obstacle-free gait, old adults exhibited a significantly more conservative strategy when crossing obstacles, with slower crossing speed, shorter step length, and shorter obstacle-heel strike distance. In addition, the old adults crossed the obstacle so that it was 10% further forward in their obstacle-crossing step. Although all subjects successfully avoided the riskiest form of obstacle contact, tripping, 4/24 healthy old adults stepped on an obstacle, demonstrating an increased risk for obstacle contact with age.

Inverse perspective mapping simplifies optical flow computation and obstacle detection

Abstract: We present a scheme for obstacle detection from optical flow which is based on strategies of biological information processing. Optical flow is established by a local "voting" (non-maximum suppression) over the outputs of correlation-type motion detectors similar to those found in the fly visual system. The computational theory of obstacle detection is discussed in terms of space-variances of the motion field. An efficient mechanism for the detection of disturbances in the expected motion field is based on "inverse perspective mapping", i.e., a coordinate transform or retinotopic mapping applied to the image. It turns out that besides obstacle detection, inverse perspective mapping has additional advantages for regularizing optical flow algorithms. Psychophysical evidence for body-scaled obstacle detection and related neurophysiological results are discussed.

Hidden Markov model for dynamic obstacle avoidance of mobile robot navigation

Abstract: Models and control strategies for dynamic obstacle avoidance in visual guidance of mobile robots are presented. Characteristics that distinguish the visual computation and motion control requirements in dynamic environments from that in static environments are discussed. Objectives of the vision and motion planning are formulated, such as finding a collision-free trajectory that takes account of any possible motions of obstacles in the local environments. Such a trajectory should be consistent with a global goal or plan of the motion and the robot should move at as high a speed as possible, subject to its kinematic constraints. A stochastic motion-control algorithm based on a hidden Markov model is developed. Obstacle motion prediction applies a probabilistic evaluation scheme. Motion planning of the robot implements a trajectory-guided parallel-search strategy in accordance with the obstacle motion prediction models. The approach simplifies the control process of robot motion. >

Motion planning in the presence of movable obstacles

Abstract: Most motion planning algorithms have dealt with motion in a static workspace, or more recently, with motion in a workspace that changes in a known manner. We consider the problem of finding collision-free motions in a changeable workspace. That is, we wish to find a motion for an object where the object is permitted to move some of the obstacles. In such a workspace, the final positions of the movable obstacles may or may not be part of the goal. In the case where the final positions of the obstacles are specified, the general problem is shown to be PSPACE-hard. In the case where the final positions of the obstacles are unspecified, the motion planning problem is shown to be NP-hard. Algorithms that run inO(n 3) time are presented for the case where there is only one movable obstacle in a polygonal environment withn corners and the object to be moved and the obstacle are convex polygons of constant complexity.

Navigational control apparatus and method for autonomus vehicles

Abstract: A navigational apparatus and method is provided for autonomous vehicles having digitally controlled drive means for driving the vehicle along a line in either forward or rearward direction and for turning the vehicle either clockwise or counterclockwise with zero turning radius. The navigational apparatus includes obstacle sensors connected to the vehicle for producing an obstacle signal in response to contact or other desired proximity between an obstacle and a vehicle as the vehicle is driven through a work area. The work area comprises a substantially horizontal planar work surface bounded on all sides by an obstacle. The navigational apparatus also includes a search memory storage device connected to the vehicle for storing a two-dimensional memory array comprising a sufficient number of array elements to represent the work area to a desired resolution and with each array element representing a unique work area increment within the work area. A data processing device connected to the vehicle communicates with the memory storage device to produce a plurality of drive signals capable of directing the vehicle's drive motors to drive the vehicle from an origin work area increment into successive work area increments until the entire work area is explored. The obstacle sensors produce obstacle signals when the vehicle encounters an obstacle in the work area and these obstacle signals are used by the data processor to navigate the vehicle around the obstacles.

Steering control system for moving vehicle

Abstract: A steering control system for a moving vehicle causes the vehicle to travel along a predetermined traveling course in a region in which a predetermined working operation, such as grass cutting or the like, is to be performed by the vehicle as it moves along the traveling course. An obstacle sensor is mounted on the vehicle, and when the sensor detects presence of an obstacle on the predetermined traveling course, the steering control system operates to cause the vehicle to leave the predetermined traveling course and to detour around the obstacle while maintaining a substantially constant distance between the obstacle and the moving vehicle. The control system is also operative to determine whether the obstacle is stationary or movable and, upon completion of the vehicle travel along the predetermined course, is operative to determine whether a movable obstacle has been removed from the predetermined course and, if so, causes the vehicle to perform its working operation in that portion of the region in which the working operation was not previously performed due to its detour around the obstacle.

On the obstacle problem for quasilinear elliptic equations of p Laplacian type

Abstract: This article considers an obstacle problem for quasilinear elliptic equations of p Laplacian type. It is shown, under certain smoothness assumptions on the obstacle, that the solution to the corresponding obstacle problem has interior Holder continuous derivatives.

Obstacle detection for high speed autonomous navigation

Abstract: Several schemes are presented for obstacle detection for autonomous vehicles traveling at high speeds (about 5 m/s). In particular, the authors discuss schemes that make a globally flat-world assumption and ignore vehicle pitch motion. They examine methods that relax the above assumptions. In each case the strengths and weaknesses of the solutions proposed are discussed. Experimental and simulation results are presented. >

Object detection in real-time

Abstract: An algorithm working on monocular gray-scale image sequences for object detectioncombined with a road tracker is presented. This algorithm, appropriate for the real-time demands of an autonomous car driving with speeds over 40 km/h, may be used for triggering obstacle avoidance maneuvers, such as coming to a safe stop automatically infront of an obstacle or following another car. Moving and static objects have been detectedin real-world experiments on various types of roads, even under unfavorable weatherconditions. Introduction . and ). A laser range finder is a usefulTo guide an autonomous road vehicle safely in _________ natural scenes (i.e rural roads, highways etc.), it is sub—task 1 necessary to recognize early enough JJ objects, (object ?) which may in some way influence the maneuvering search of the vehicle. Such objects might be, for instance, detect vehicles or obstacles in front of the car, passing or alarm crossing vehicles, or traffic signs. This paper con-centrates on the problem of detecting new objects,

A Comparison of Real-Time Obstacle Avoidance Methods for Mobile Robots

Abstract: Recently several new dynamic approaches have been developed to enable a vehicle to autonomously avoid obstacles in its environment in real time. These algorithms are generally considered as reflexive collision avoidance algorithms since they are continuously using the latest update in sensory data and computing from this data error signals to drive and/or steer the vehicle away from a collision with the environment. This paper experimentally tested three of these methods, potential fields, generalized potential fields and vector field histograms, using a uniform set of hardware and software modules. The types of tests chosen were typical of navigation in an indoor environment and consisted of avoiding a single obstacle at high speed, travelling through a narrow hallway and passing through an open doorway. The main issues observed during the testing were, the maximum speed at which the vehicle could accomplish the test, the nature of the path taken by the vehicle during the test and any difficulties that arose in the process of implementing any of the algorithms.

Coping with uncertainty in control and planning for a mobile robot

Abstract: Describes a decision theoretic approach to real-time obstacle avoidance and path planning for a mobile robot. The mobile robot navigates in a semi-structured environment in which unexpected obstacles may appear at random locations. Twelve sonar sensors are currently used to report the presence and location of the obstacles. To handle the uncertainty of an obstacle's appearance, the authors adopt a Bayesian approach by assuming a prior distribution for the presence of unknown obstacles. The distribution is changed dynamically according to the information accumulated by sensors. When searching for an optimal path using dynamic programming, the authors take the probability into account in making a decision. Based on prior information and sensor data, they show that the proposed method allows the mobile robot to avoid unexpected obstacles and finds an optimal path to the goal in real time. >

A ground and obstacle collision avoidance technique (GOCAT)

Abstract: The authors present a novel solution to the problem of providing aircrew with a timely warning of a dangerously close approach to the terrain. An overview of why such a capability is required, the benefits it is expected to provide, and specific requirements that it must fulfil is given. The ground proximity warning techniques which are currently available are reviewed, and the benefits and limitations of each technique are assessed. The GEC Avionics Ground and Obstacle Collision Avoidance Technique (GOCAT), which forms part of the Total Terrain Avionics system, is introduced. The particular advantages of the GOCAT approach, such as the use of a terrain database, are identified, and issues of parameter selection, search area definition, and system limitations are described. The results of a GOCAT simulation are presented to demonstrate the conditions under which a warning is required and generated. The characteristics of a practical implementation for both civil and military applications are discussed. >

Visual navigation around curved obstacles

Abstract: An approach to path-planning around smooth obstacles that exploits visually derived geometry is proposed. A moving robot can scan the silhouette or apparent contour of an obstacle and estimate a minimum length path. This is done by seeking geodesics which can be extrapolated smoothly, around the obstacle and towards the goal. Preliminary implementation of this idea uses a real-time visual contour tracker running at 16 Hz, with a camera mounted on an Adept robot arm. The camera first dithers to generate visual motion, a safe path is estimated, and the robot steers the camera around the obstacle with a clearance of a few millimeters. >

Considerations for Automated Nap-of-the-Earth Rotorcraft Flight

Abstract: In this paper, we consider nap-of-the-earth (NOE) rotorcraft flight as one of the applications in which obstacle avoidance plays a key role, and investigate the prospects of automating the guidance functions of NOE flight. Based on a proposed structure for the guidance functions, we identify obstacle detection and obstacle avoidance as the two critical components requiring substantial advancement before an automatic guidance system can be realized. We discuss the major sources of difficulties in developing these two components, including sensor requirements for which we provide a systematic analysis.

Object classification for obstacle avoidance

Abstract: Object recognition is necessary for any mobile robot operating autonomously in the realworld. This paper discusses an object classifier based on a 2-D object model. Obstaclecandidates are tracked and analyzed, false alarms generated by the object detector arerecognized and rejected. The methods have been implemented on a multi-processorsystem and tested in real-world experiments. They work reliably under favorableconditions, but sometimes problems occur, e.g. when objects contain many features(edges) or move in front of structured background. Introduction An autonomous vehicle participating in road traffichas to master various traffic situations. Avoiding acollision with a moving or static object is an impor-tant subtask. Such objects will be called obstaclesin the sequel, regardless of their nature. Obstaclesmay appear in the environment of the autonomousvehicle at any time. Due to reasons of safety, it isnecessary to reliably detect, locate and analyze allobstacles without exception. In spite of the requiredhigh reliability each of these processes must be per-formed quickly, because the maximum permissiblespeed of the vehicle depends on the time requiredfor recognizing an obstacle and initiating an avoid-

Sound field in a rectangular cavity in the presence of a thin, flexible obstacle by the integral equation method

Abstract: This paper is concerned with the prediction of the sound field in a cavity enclosing a thin, flexible obstacle and having finite impedance boundaries. The formulation uses the integral equation method adapted to consider a thin obstacle. It also makes use of the equation of motion of the obstacle, but the only form given here is for the case of a limp panel. The solution is achieved by expressing all the required functions as modal expansions in the equations defining the problem. The set of eigenfunctions of an empty, rigid-walled rectangular cavity is used as a basis for the expansions, which restricts the final solution to rectangular cavities. A few numerical results are presented to validate the theory on some aspects related to simple cases of plane obstacles in rigid-walled cavities.

Obstacle alarm device

Abstract: PURPOSE:To provide an obstacle alarm device which surely detects the existence and position of an obstacle even if an occupant does not pay attention to the obstacle alarm device. CONSTITUTION:An obstacle detecting means 1 detects an obstacle within a prescribed range on the periphery of a vehicle. And a sound control means 3 controls the sound generated by a sound device 2, according to the existence direction of the obstacle.

A System for Fast Navigation of Autonomous Vehicles

Abstract: : This report describes an autonomous mobile robot designed to navigate at high speeds over featureless terrain-- terrain which is not navigable by relying on features that are found on paved roads and highways. To this end we have developed a paradigm that we call position based navigation that relies on explicit vehicle position information from inertial and satellite instruments, to navigate. Specifically we have tackled four main areas--path tracking, which guides the robot vehicle over a pre-specified path; obstacle detection, which is responsible for bringing the vehicle to a stop from high speed when an obstacle is detected; obstacle avoidance, which is responsible for steering the vehicle around detected obstacles so as to rejoin the specific path; and computing architecture, which integrates all the capabilities into one system. We discuss the algorithms and the devices that were implemented on NavLab, a navigation testbed at Carnegie Mellon. The most notable of our results is the 11m/s speed achieved by the vehicle.

Cascading a systolic array and a feedforward neural network for navigation and obstacle avoidance using potential fields

Abstract: A technique is developed for vehicle navigation and control in the presence of obstacles. A potential function was devised that peaks at the surface of obstacles and has its minimum at the proper vehicle destination. This function is computed using a systolic array and is guaranteed not to have local minima. A feedfoward neural network is then used to control the steering of the vehicle using local potential field information. In this case, the vehicle is a trailer truck backing up. Previous work has demonstrated the capability of a neural network to control steering of such a trailer truck backing to a loading platform, but without obstacles. Now, the neural network was able to learn to navigate a trailer truck around obstacles while backing toward its destination. The network is trained in an obstacle free space to follow the negative gradient of the field, after which the network is able to control and navigate the truck to its target destination in a space of obstacles which may be stationary or movable.

Method for estimating position of concerning obstacle from vehicle

Abstract: PURPOSE: To improve the position estimating accuracy of a concerning obstacle by estimating the position of the obstacle after a prescribed period of time has elapsed based on the relative velocity vector of each obstacle against its own vehicle. CONSTITUTION: After the position information of the concerning obstacle from a distance measuring unit 32 is expanded on an X (width direction of vehicle) and Y (length direction of vehicle) coordinates, a predictive arithmetic section 51 compares the current data with predicted data and outputs the data which are discriminated as the data of the same obstacle from a comparing arithmetic section 52 and a relative velocity vector calculating section 53 calculates the relative velocity vector of the objective obstacle. The calculated vector is inputted to obstacle position calculating sections 54-57 and the outputs of the sections 54 and 55 and 56 and 57 are respectively inputted to a discriminating means 40 and control means 41 in accordance with the output signal of an awakening-degree detecting section 37. On the other hand, the collision of its own vehicle with the obstacle is avoided or, when the collision is unavoidable, the damage of the vehicle is reduced by estimating the position of its own vehicle after set time has elapsed from the braking starting time by means of an its-own-vehicle position estimating means 38 and actuating an alarm S 2 or brake A 1 by comparing the estimated positions of the obstacle and the vehicle. COPYRIGHT: (C)1993,JPO&Japio

The variable modelling of mobile robot environments

Abstract: The modelling of a mobile robot environment is mainly motivated first to obtain a map of the free space (in order to compute safe trajectories for the vehicle) and second to provide the control loop with feed-back of obstacle information. Here, the authors only consider the problem of local representations for the displacements of a vehicle with two preoccupations: the environment models have to be adequate for the robot navigation algorithms, and the construction of these models have to be as fast as possible. The study is carried out at two levels: a low level based on the analysis of the free zones of the environment and a high level of obstacle modelling. According to local navigation, the obstacle modelling is based on the artificial potential field method. >

Self-running work robot

Abstract: PURPOSE:To automatically attain work on the place which has not been worked yet by storing the position of an obstacle detected by an obstacle sensor, and after evading the obstacle and ending a planned operation, moving a robot to the stored position and making it execute the residual work at the spot if the obstacle is removed. CONSTITUTION:Responding to an obstacle detection signal (a) by an ultrasonic sensor 26, a position information of a robot at the time of detecting an obstacle is supplied from a position advancing direction calculating part 13 to an obstacle position comparing part 23. When an obstacle is detected on a travelling course, the robot detours an obstacle and the operation is continued. And after the general operation is ended, the presence of the obstacle can be rechecked by returning the robot 1 to the position where, in the obstacles being detected, obstacles other than the ones which are preliminarily registered as fixed obstacles, and ones which were registered as a result of a learning process, are detected. Thus, if the obstacle is moved, the residual work can be executed at the place.

Obstacle avoidance using hierarchical dynamic programming

Abstract: Studies the problem of finding the optimal path or trajectory of a robot or other unmanned device through an obstacle field for a given start and goal position. In the most general problem, the obstacle field would be modeled by a finite array of time dependent analog valued pixels. Moreover since the field of vision of the unmanned device is limited, the authors could not assume perfect knowledge of the full obstacle field at any given time. The author discuss approximate optimal paths constructed using hierarchical methods. The hierarchical methods entail constructing a coarse resolution version of the original obstacle array and then using multipass dynamic programming to find an optimal coarse path. >

Obstacle evading method for unmanned carrier

Abstract: PURPOSE:To simply evade the obstacles for an unmanned vehicle by translating the reference coordinate axes away from the obstacles and defining the shifted coordinate axes as an obstacle evading source. CONSTITUTION:An unmanned carrier 11 is provided with a drive controller having a prescribed arithmetic/storage function which guides the carrier 11 along a driving course 10. Then the carrier 11 inputs the signals received from various sensors like a guidance detection coil, a steering angle detecting potentiometer of a steering/driving wheel 12, the encoders provided at the right and left fixed wheels 12a and 12b, etc. Then the position of an obstacle is computed on a carrier body fixed coordinate axes based on the detection signal of an obstacle sensor. At the same time, this position of the obstacle is transformed into a position on a reference coordinate axes. Then the reference coordinate axes are translated in the direction so as to be away from the obstacle, and the driving/steering control is carried out on the new coordinate axes. In such a constitution, the obstacles can be simply evaded with driving/steering operations performed in a short time and without increasing the constitutional scale.

Using Inverse Perspective Mapping as a Basis for two Concurrent Obstacle Avoidance Schemes

Abstract: We describe two real-time visual obstacle avoidance schemes based on a common preliminary co-ordinate transformation called Inverse Perspective Mapping (IPM) [1]. The first scheme makes use of motion discontinuities in the mapped motion field due to the obstacle. An alternative, stereoptical approach extracts the obstacle by discriminating deviations in the iso-disparity surfaces of the transformed image pair. Both concurrent schemes are implemented on a mobile robot.

Status of automated nap-of-the-earth rotorcraft guidance

Abstract: The status of automated nap-of-the-earth (NOE) flight guidance is discussed by examining two current research efforts that approach the problem of performing obstacle avoidance along a pre-planned course from different perspectives. The first, a real-time guidance system developed by Systems Technology Inc., focuses on the issue of pilot acceptable maneuvers, under the assumption of an ideal, but nonrealizable, obstacle detection system. The second, a system created at NASA Ames, concentrates on the utilization of realistic on-board sensor data for obstacle detection and avoidance-maneuver decisions. This paper examines the strengths and weaknesses of these two approaches, and how they might complement each other. The goal is an automatic NOE system that is both acceptable to pilots and capable of making intelligent use of realistic sensor information and path-selection logic.