室内移动机器人视觉里程计研究

研究了一种精度较高的视觉里程计实现方法。首先采用计算量小、精度高的SURF算法提取特征点,并采用最近邻向量匹配法进行特征点匹配;采用RANSAC算法求得帧间摄像机坐标系的转换矩阵;提出了一种视觉里程计标定方法,并通过该方法得到机器人起始点坐标系下的位移信息,保证了机器人有自身姿态变化时的定位的准确性;圆轨迹实验验证了视觉里程计算法的可行性。     

轮式移动机器人里程计系统误差校核*

里程计使用编码器为轮式移动机器人提供基本的位姿估计,在运行过程中里程计存在严重的误差累计,通过校核系统参数可以减小系统误差,UMBmark方法是轮式移动机器人广泛使用的系统误差校核方法。针对UMBmark方法存在的不足,提出一种改进的系统误差校核新方法：综合考虑三种主要系统误差来源产生的误差对移动机器人直线运动和定点旋转运动造成的影响,同时采用正方形回路终点的方向误差代替传统UMBmark方法中的位置误差来校核系统参数。实验结果表明提出的方法能够有效校核系统参数,提高移动机器人的定位精度。     

VISYR: a simulation tool for mobile robots using vision sensors

With the advent of mobile robots and inboard vision sensors mounted directly on the robot's wrist, new kinds of problems lie in the image processing field as, for example, dynamic scene analysis or motion estimation. The lack of flexibility of real experiments led us to implement at IRISA a general simulation tool devoted to the study of robots using moving vision sensors. VISYR allows us to simulate the image perceived by a robot of its environment during its motion.The first part of the paper is devoted to the modelling of the 3D scene containing complex objects and to the design of a suitable robotics vision sensor. In the second part, a new algorithm of dynamic management of the local data basis perceived by the sensor is presented. The parameters of the vision sensors are highly adjustable and VISYR is conceived to allow the fast development of algorithms using dynamic vision data.     

A calibration method for odometry of mobile robots based on the least-squares technique: theory and experimental validation

For a mobile robot, odometry calibration consists of the identification of a set of kinematic parameters that allow reconstructing the vehicle's absolute position and orientation starting from the wheels' encoder measurements. This paper develops a systematic method for odometry calibration of differential-drive mobile robots. As a first step, the kinematic equations are written so as to underline linearity in a suitable set of unknown parameters; thus, the least-squares method can be applied to estimate them. A major advantage of the adopted formulation is that it provides a quantitative measure of the optimality of a test motion; this can be exploited to drive guidelines on the choice of the test trajectories and to evaluate accuracy of a solution. The proposed technique has been experimentally validated on two different mobile robots and, in one case, compared with other existing approaches; the obtained results confirm the effectiveness of the proposed calibration method.     

Using infrared sensors for distance measurement in mobile robots

The amplitude response of infrared (IR) sensors based on reflected amplitude of the surrounding objects is non-linear and depends on the reflectance characteristics of the object surface. As a result, the main use of IR sensors in robotics is for obstacle avoidance. Nevertheless, their inherently fast response is very attractive for enhancing the real-time operation of a mobile robot in, for instance, map building tasks. Thus, it seems that the development of new low-cost IR sensors able to accurately measure distances with reduced response times is worth researching. In this paper, a new IR sensor based on the light intensity back-scattered from objects and able to measure distances of up to 1 m is described. Also, the sensor model is described and the expected errors in distance estimates are analysed and modelled. Finally, the experimental results obtained are discussed.     

A method for combining odometry and distance sensor information for effective obstacle avoidance of autonomous mobile robots

In this paper, a concept for virtual sensors is proposed for efficient avoidance of obstacles during the motion of robots. The virtual sensor yields new data by combining encoder values and real distance data, and derives new sensor data that includes the mobility of the robot. Simulation on Windows XP is executed to illustrate the proposed approach with actually acquired distance from virtual and actual sensors. To facilitate comparison with the alternative results developed in this paper, we refer to the conventional artificial potential field method using actual distance. Data from virtual sensors show smoother and safer motion in obstacle avoidance traces in regards to obstacle and robot mobility.     

Motion planning for mobile robots using a laser range finder

We have designed a mobile robot with a distribution structure for intelligent life space. The mobile robot was constructed using an aluminum frame. The mobile robot has the shape of a cylinder, and its diameter, height, and weight are 40 cm, 80 cm, and 40 kg, respectively. There are six systems in the mobile robot, including structure, an obstacle avoidance and driving system, a software development system, a detection module system, a remote supervision system, and others. In the obstacle avoidance and driving system, we use an NI motion control card to drive two DC servomotors in the mobile robot, and detect obstacles using a laser range finder and a laser positioning system. Finally, we control the mobile robot using an NI motion control card and a MAXON driver according to the programmed trajectory. The mobile robot can avoid obstacles using the laser range finder, and follow the programmed trajectory. We developed the user interface with four functions for the mobile robot. In the security system, we designed module-based security devices to detect dangerous events and transmit the detection results to the mobile robot using a wireless RF interface. The mobile robot can move to the event position using the laser positioning system.     

基于反步法的移动机器人鲁棒跟踪控制

以四轮移动机器人为研究对象,建立了机器人完整的数学模型,包括运动学模型、动力学模型以及驱动电机模型。在机器人数学模型的基础上,采用反步法的思想设计具有全局收敛特性的鲁棒轨迹跟踪控制器,设计中考虑了驱动电机模型使控制器更符合实际控制要求,并将其分解为运动学控制器、动力学控制器以及电机控制器三部分,降低了控制器设计的难度。构造了系统的李雅普诺夫函数,证明了该类型移动机器人在所得控制器作用下,能实现对给定轨迹的全局渐近追踪。仿真实验结果表明基于反步法的控制器是有效的。     

Path following with a security margin for mobile robots

This paper addresses the problem of determining a feedback control law, robust with respect to localization errors, allowing a mobile robot to follow a prescribed path. The model that we consider is a dynamic extension of the usual kinematic model of a mobile robot in the sense that the path curvature is defined as a new state variable. The control variables are the linear velocity and the derivative of the curvature. By defining a sliding manifold we determine a stabilizing controller for the nominal system, that is when the exact configuration is supposed to be known. Then, we prove that the system remains stable when the feedback control inputs use estimated values instead of the exact values, and we characterize the control robustness with respect to localization and curvature estimation errors. The control robustness is expressed by determining a bounded attractive domain containing the configuration error as the closed-loop control is performed with the estimated state values. Two control laws are successively proposed. The former is deduced from Lyapunov's direct method, and the latter is based on variable structure control techniques. Using variable structure control we show that the size of the attractive domain can be easily minimized while keeping the balance between short response time, low output oscillation, and large stability domain. Knowledge of this attractive domain allows us to compute easily a security margin to guarantee obstacle avoidance during the path following process. Experimental results are presented at the end of the paper.     

Attention-based navigation in mobile robots using a reconfigurable sensor

In this paper, a method for visual attentional selection in mobile robots is proposed, based on amplification of the selected stimulus. Attention processing is performed on the vision sensor, which is integrated on a silicon chip and consists of a contrast sensitive retina with the ability to change the local inhibitory strength between adjacent pixel elements. As a result, the sensitivity to visual contrast at a particular region of the retina can be adjusted. As the local inhibitory strength can be regulated from outside of the chip, a reconfigurable sensor is realized. This “attention-retina” was tested on an autonomous robot (MorphoII) which was given the task of selecting a line to follow while there were two alternatives. The robot develops directional preference by associating its visual stimulus with an electrical energy providing stimulus, in this case a solar cell.     

Learning control of mobile robots using a multiprocessor system

A real-time multiprocessor system is proposed for the solution of the tracking problem of mobile robots operating in a real context with environmental disturbances and parameter uncertainties. The proposed control scheme utilizes multiple models of the robot for its identification in an adaptive and learning control framework. Radial Basis Function Networks (RBFNs) are considered for the multiple models in order to exploit the net non-linear approximation capabilities for modeling the kinematic behavior of the vehicle and for reducing unmodeled contributions to tracking errors. The training of the nets and the tests of the achieved control performance have been done in a real experimental setup. The proposed control architecture improves the robot tracking performance achieving fast and accurate control actions in presence of large and time-varying uncertainties in dynamical environments. The experimental results are satisfactory in terms of tracking errors and computational efforts.     

Mobility characterization for autonomous mobile robots using machine learning

This paper presents a supervised learning approach to improving the autonomous mobility of wheeled robots through sensing the robot’s interaction with terrain ‘underfoot.’ Mobility characterization is cast as a hierarchical task, in which pre-immobilization detection is achieved using support vector machines in time to prevent full immobilization, and if a pre-immobilization condition is detected, the associated terrain feature affecting mobility is identified using a Hidden Markov model. These methods are implemented using a hierarchical, layered control scheme developed for the Yeti robot, a 73-kg, four-wheeled robot designed to perform autonomous medium-range missions in polar terrain. The methodology is motivated by the difficultly of visually recognizing terrain features that impact mobility in low contrast terrain. The efficacy of the approach is evaluated using data from a suite of proprioceptive sensors. Real-time implementation shows that Yeti can consistently detect pre-immobilization conditions, stop in time to avoid unrecoverable immobilization, identify the terrain feature presenting the mobility challenge, and execute an escape sequence to retreat from the condition.     

Real-time people tracking for mobile robots using thermal vision

This paper presents a vision-based approach for tracking people on a mobile robot using thermal images. The approach combines a particle filter with two alternative measurement models that are suitable for real-time tracking. With this approach a person can be detected independently from current light conditions and in situations where no skin colour is visible. In addition, the paper presents a comprehensive, quantitative evaluation of the different methods on a mobile robot in an office environment, for both single and multiple persons. The results show that the measurement model that was learned from local grey-scale features could improve on the performance of the elliptic contour model, and that both models could be combined to further improve performance with minimal extra computational cost.     

Fault detection for service mobile robots using model-based method

Detection of faults is a topic of high importance because it increases robot dependability, a requirement for the wide acceptance of service robots in domestic environments. This work takes a model-based approach for detecting and identifying actuator faults on differential-drive mobile robots in an indoor environment. An error-bound is calculated between the estimated and measured robot states which is constantly adapted based on the current state and input signals. A fault is detected when the estimation error is outside this bound. The model parameters are learned by the robot using an adaptive law, after the robot deployment in the target environment. Model uncertainties have an important impact on the fault detection performance, and are dealt with by considering the uncertainty bounds in the bound calculations. This ensures no false alarms occur when the uncertainty remains bounded during normal operation. Furthermore an extension to the method is proposed that addresses the problem of detecting small faults. The method is experimentally validated on a iRobot Roomba autonomous robot.     

Global localization for mobile robots using reference scan matching

This paper presents a new approach based on scan matching for global localization with a metric-topological hybrid world model. The proposed method aims to estimate relative pose to the most likely reference site by matching an input scan with reference scans, in which topological nodes are used as reference sites for pose hypotheses. In order to perform scan matching we apply the spectral scan matching (SSM) method that utilizes pairwise geometric relationships (PGR) formed by fully interconnected scan points. The SSM method allows the robot to achieve scan matching without using an initial alignment between two scans and geometric features such as corners, curves, or lines. The localization process is composed of two stages: coarse localization and fine localization. Coarse localization with 2D geometric histogram constructed from the PGR is fast, but not precise sufficiently. On the other hand, fine localization using the SSM method is comparatively slow, but more accurate. This coarse-to-fine framework reduces the computational cost, and makes the localization process reliable. The feasibility of the proposed methods is demonstrated by results of simulations and experiments.     

Appearance-based mapping and localization for mobile robots using a feature stability histogram

The strength of appearance-based mapping models for mobile robots lies in their ability to represent the environment through high-level image features and to provide human-readable information. However, developing a mapping and a localization method using these kinds of models is very challenging, especially if robots must deal with long-term mapping, localization, navigation, occlusions, and dynamic environments. In other words, the mobile robot has to deal with environmental appearance change, which modifies its representation of the environment. This paper proposes an indoor appearance-based mapping and a localization method for mobile robots based on the human memory model, which was used to build a Feature Stability Histogram (FSH) at each node in the robot topological map. This FSH registers local feature stability over time through a voting scheme, and the most stable features were considered for mapping, for Bayesian localization and for incrementally updating the current appearance reference view in the topological map. The experimental results are presented using an omnidirectional images dataset acquired over the long-term and considering: illumination changes (time of day, different seasons), occlusions, random removal of features, and perceptual aliasing. The results include a comparison with the approach proposed by Dayoub and Duckett (2008) [19] and the popular Bag-of-Words (Bazeille and Filliat, 2010) [35] approach. The obtained results confirm the viability of our method and indicate that it can adapt the internal map representation over time to localize the robot both globally and locally.     

Detection of doors using a genetic visual fuzzy system for mobile robots

Doors are common objects in indoor environments and their detection can be used in robotic tasks such as map-building, navigation and positioning. This work presents a new approach to door-detection in indoor environments using computer vision. Doors are found in gray-level images by detecting the borders of their architraves. A variation of the Hough Transform is used in order to extract the segments in the image after applying the Canny edge detector. Features like length, direction, or distance between segments are used by a fuzzy system to analyze whether the relationship between them reveals the existence of doors. The system has been designed to detect rectangular doors typical of many indoor environments by the use of expert knowledge. Besides, a tuning mechanism based on a genetic algorithm is proposed to improve the performance of the system according to the particularities of the environment in which it is going to be employed. A large database of images containing doors of our building, seen from different angles and distances, has been created to test the performance of the system before and after the tuning process. The system has shown the ability to detect rectangular doors under heavy perspective deformations and it is fast enough to be used for real-time applications in a mobile robot.     

Self-calibration of gyro using monocular SLAM for an indoor mobile robot

Rate gyros are widely used to calculate the heading angle for mobile robot localization. They are normally calibrated in the factory using an expensive rate table prior to their use. In this paper, a self-calibration method using a monocular camera without a rate table is proposed. The suggested method saves time and cost for extra calibration procedure. SLAM (Simultaneous Localization And Mapping) based on visual features and odometry gives reference heading (yaw) angles. Using these, the coefficients of a scale factor function are estimated through Kalman filtering. A new undelayed feature initialization method is proposed to estimate the heading angle without any delay. Experimental results show the efficiency of the proposed method.