Terrain adaptive detector selection for visual odometry in natural scenes

Pose estimation is one of the important tasks for mobile robots exploring in outdoor environments. Recently, visual odometry has received a lot of attention since its localization is accurate even with low-cost sensors. Furthermore, the technique is not affected by wheel slips, and it can be performed without external infrastructures and preliminary maps. While existing techniques successfully provide good localization of outdoor vehicles, possible failures are not yet fully examined in untextured terrains where feature tracking occasionally fails. This paper proposes an approach to detect interest points from a wide variety of terrains by adaptively selecting algorithms. Experiments show that the approach provides robust and fast interest point detection even in untextured natural scenes.     

Estimation of Terrain Forces and Parameters for Rigid-Wheeled Vehicles

This paper provides a methodology for the estimation of resistance, thrust, and resistive torques on each wheel of a rigid-wheeled vehicle generated at the vehicle–terrain interface, and from these forces and moments, a methodology to estimate terrain parameters is presented. Terrain force estimation, which is independent of a terrain model, can infer the ability to accelerate, climb, or tow a load independent of the underlying terrain properties. When a terrain model is available, parameters of that model, such as soil cohesion, friction angle, maximum normal stress, and stress distribution parameters, are determined from estimated vehicle–terrain forces using a multiple-model estimation approach, providing parameters that relate to accepted mobility metrics. The methodology requires a standard proprioceptive sensor suite—accelerometers, rate gyros, wheel speeds, motor torques, and ground speed. Sinkage sensors are not required. Simulation results demonstrate efficacy of the method on three terrains spanning a range of soil cohesions reported in the literature.      

Game-based pursuit evasion for nonholonomic wheeled mobile robots subject to wheel slips

Pursuit-evasion (P-E) problem has been studied as a noncooperative zero-sum game in homicidal chauffer problem in 1960s and in the game of two identical cars most recently. The capture conditions in the two games, which govern the capture behavior, can be determined by solving Hamilton–Jacobi–Isaacs equations. However, the existing game theoretic solution does not consider wheel slip, and consequently, cannot answer the escape and capture conditions in the presence of wheel slip. In this paper, we investigate how to predict capture and escape conditions when the pursuer has wheel slip. We study a dynamic P-E game problem with a nonholonomic wheeled mobile robot (WMR) pursuer subject to wheel slip and propose an equivalent kinematic model to develop escape and capture conditions in the presence of wheel slip. To our knowledge, this is the first time the P-E game problems with WMR have been analyzed with wheel slip. The presented framework will allow future development of realistic P-E strategies that do not ignore wheel slip and thus will be able to model high-speed P-E on different terrains.     

State estimation and traversability map construction method of a quadruped robot on soft uneven terrain

Quadruped robots show excellent application prospects in complex environment detection and rescue. At present, scholars mainly focus on quadruped walking in rigid environments. However, quadruped robots often need to pass through uneven and soft unconstructed terrains, prone to slip and impact. The mismatch between the planned foothold position and the real one resulting from environmental uncertainties makes the robot unstable. In this paper, the state estimation and traversability map construction methods are proposed for quadruped robots to achieve stable walking in an unstructured environment, especially on soft terrains. First, the Error-state Kalman Filter (ErKF) is extended by optimizing the leg odometry information to get an accurate robot state, especially in soft, uneven terrain. The ErKF method fuses the sensor data from the inertial measurement unit, laser, camera, and leg odometry. The leg odometry is optimized by considering the foot slippage, which easily occurs in soft uneven terrains. Then, the unstructured environment is parameterized and modeled by the terrain inclination, roughness, height, and stiffness. A traversability map, which is essential for robot path and foothold planning in autonomous movement, is constructed with the above parameters. Finally, the proposed method is verified by simulation and experiments. The results show that the quadruped robot can walk stably on different soft and uneven terrains.     

Positioning device for outdoor mobile robots using optical sensors and lasers

We propose a novel method for positioning a mobile robot in an outdoor environment using lasers and optical sensors. Position estimation via a noncontact optical method is useful because the information from the wheel odometer and the global positioning system in a mobile robot is unreliable in some situations. Contact optical sensors such as computer mouse are designed to be in contact with a surface and do not function well in strong ambient light conditions. To mitigate the challenges of an outdoor environment, we developed an optical device with a bandpass filter and a pipe to restrict solar light and to detect translation. The use of two devices enables sensing of the mobile robot’s position, including posture. Furthermore, employing a collimated laser beam allows measurements against a surface to be invariable with the distance to the surface. In this paper, we describe motion estimation, device configurations, and several tests for performance evaluation. We also present the experimental positioning results from a vehicle equipped with our optical device on an outdoor path. Finally, we discuss an improvement in postural accuracy by combining an optical device with precise gyroscopes.     

A fusion strategy for reliable vehicle positioning utilizing RFID and in-vehicle sensors

In recent years, RFID has become a viable solution to provide object's location information. However, the RFID-based positioning algorithms in the literature have disadvantages such as low accuracy, low output frequency and the lack of speed or attitude information. To overcome these problems, this paper proposes a RFID/in-vehicle sensors fusion strategy for vehicle positioning in completely GPS-denied environments such as tunnels. The low-cost in-vehicle sensors including electronic compass and wheel speed sensors are introduced to be fused with RFID. The strategy adopts a two-step approach, i.e., the calculation of the distances between the RFID tags and the reader, and then the global fusion estimation of vehicle position. First, a Least Square Support Vector Machine (LSSVM) algorithm is developed to obtain the distances. Further, a novel LSSVM Multiple Model (LMM) algorithm is designed to fuse the data obtained from RFID and in-vehicle sensors. Contrarily to other multiple model algorithms, the LMM is more suitable for current driving conditions because the model probabilities can be calculated according to the operating state of the vehicle by using the LSSVM decision model. Finally, the proposed strategy is evaluated through experiments. The results validate the feasibility and effectiveness of the proposed strategy.     

A versatile locomotion mechanism for amphibious robots: eccentric paddle mechanism

Most of recently developed rescue robots can only be deployed to limited attacked regions after tsunami and the floods, due to their limited mobility on complex amphibious terrains. To access such amphibious environments with improved mobility, we propose a novel eccentric paddle mechanism (ePaddle) which has a set of paddles eccentrically placed in a wheel to perform multiple terrestrial, aquatic, and amphibious gaits. One of the advantages of our proposed ePaddle mechanism is its unique locomotion versatility introduced by the eccentric distance between the paddle shaft and the wheel center. We demonstrate this versatility by proposing five typical gaits for traveling on different terrains. For instance, wheeled rolling gait is used to achieve high-speed locomotion on even terrain. Legged gait is applied to travel on the rough terrains. To access the soft terrains where wheels slip and legs sink, a wheel-leg-integrated gait is performed by digging the paddle into the ground. To swim in the water, rotational paddling and oscillating paddling gaits are proposed. For each of these gaits, standard gait sequence is defined and joint parameters are calculated based on kinematics. An ePaddle prototype is then built and tested with the proposed gait sequences. Experimental results verify the design of the ePaddle mechanism as well as its versatile gaits.     

Slippage and immobilization detection for planetary exploration rovers via machine learning and proprioceptive sensing

This paper presents a new methodology where machine learning is used for detecting various levels of slip in the context of planetary exploration robotic missions. This methodology aims at employing proprioceptive rover sensor signals. Consequently, no operational complexity is added to the rover's commanding and it is independent of lighting conditions. Two supervised learning methods (Support Vector Machines and Artificial Neural Networks) are compared to two unsupervised learning approaches (K‐means and Self‐Organizing Maps (SOM)). Physical experiments using a single‐wheel testbed equipped with an MSL spare wheel and a real planetary exploration rover validate the implemented methodology. Performance is evaluated in terms of well‐known metrics both considering single data points and subsets of consecutive data points (moving median filter). Computation time and storage requirements are also examined. One of the SOM‐based algorithms, semantic SOM method, demonstrates a proper balance between the benefits of supervised learning algorithms (high success rate, >96%) and the advantages of unsupervised learning methods (low storage requirements, 5 kb, and no need of manually‐labeled training data). This paper also addresses the most convenient placement of IMU sensors on the rover chassis such that slippage detection is maximized.     

Mars Science Laboratory Curiosity Rover Megaripple Crossings up to Sol 710 in Gale Crater

After landing in Gale Crater on August 6, 2012, the Mars Science Laboratory Curiosity rover traveled across regolith‐covered, rock‐strewn plains that transitioned into terrains that have been variably eroded, with valleys partially filled with windblown sands, and intervening plateaus capped by well‐cemented sandstones that have been fractured and shaped by wind into outcrops with numerous sharp rock surfaces. Wheel punctures and tears caused by sharp rocks while traversing the plateaus led to directing the rover to traverse in valleys where sands would cushion wheel loads. This required driving across a megaripple (windblown, sand‐sized deposit covered by coarser grains) that straddles a narrow gap and several extensive megaripple deposits that accumulated in low portions of valleys. Traverses across megaripple deposits led to mobility difficulties, with sinkage values up to approximately 30% of the 0.50 m wheel diameter, resultant high compaction resistances, and rover‐based slip up to 77%. Analysis of imaging and engineering data collected during traverses across megaripples for the first 710 sols (Mars days) of the mission, laboratory‐based single‐wheel soil experiments, full‐scale rover tests at the Dumont Dunes, Mojave Desert, California, and numerical simulations show that a combination of material properties and megaripple geometries explain the high wheel sinkage and slip events. Extensive megaripple deposits have subsequently been avoided and instead traverses have been implemented across terrains covered with regolith or thin windblown sand covers and megaripples separated by bedrock exposures.     

基于UWB-PDOA的少基站自适应定位系统研究

超宽带(Ultra-Wideband,UWB)技术在室内外定位中应用广泛,针对传统多基站定位方案的局限性,提出了一种基于超宽带信号到达相位差(Ultra-Wideband Phase Difference of Arrival,UWB-PDOA)的少基站自适应定位系统。该系统利用UWB-PDOA技术和基于ESP32信号强度的权重自适应定位技术,大幅降低了对环境部署的依赖性,提高了定位的精度和稳定性。结合环境先验信息和目标高度的先验知识,构建了先验知识库,采用自适应定位技术,利用多个传感器的信息来调整对不同定位基站的置信度权重,进一步提高了定位精度和鲁棒性。实验结果表明,所提出的系统在视距(Line of Sight,LOS)和非视距(Non Line of Sight,NLOS)环境下都具有较高的定位精度和稳定性,并且仅需要不超过3个基站便可以满足室内环境定位的需求。     

Online Learning and Adaptation of Autonomous Mobile Robots for Sustainable Agriculture

In this paper we will introduce the application of our newly patented double hierarchical Fuzzy-Genetic system (British patent 99-10539.7) to produce an intelligent autonomous outdoor agricultural mobile robot capable of learning and calibrating its controller online in a short time interval and implementing a life long learning strategy. The online and life long learning strategy allow the outdoor robots to increase their experience and adapt their controllers in the face of the changing and dynamic unstructured outdoor agricultural environments. Such characteristics permit prolonged periods of operation within dynamic agricultural environments, which is an essential feature for the realization of a platform vehicle for use in sustainable agriculture and organic farming.     

室外崎岖地形下基于视差图的无人自主车障碍物识别

对于地面无人车和室外非结构化环境, 本文介绍了我们开发的基于立体视觉的障碍物快速识别系统. 为了使无人地面车适应于较复杂的地形, 根据V视差图, 我们提出了一种新的地面主视差图的估计方法. 通过地面主视差图和局部的三维重建, 本文给出了一种由粗到精的障碍物识别与定位方法. 在我们的无人地面车平台上, 我们对这一障碍物自动识别系统进行了相应的实际试验. 其试验结果验证了该系统的有效性.     

Sensor Fusion System Using Recurrent Fuzzy Inference

In robotic and manufacturing systems, it is difficult to measure the state of systems accurately because of many uncertain factors and noise, and it is very important to estimate the state of systems. We must measure the phenomena of systems by multiple sensors and estimate the state of systems by acquiring information of sensors. However, we can not acquire all of sensor information synchronically, because each sensor has particular sensor information and measuring time. For estimating the state of systems by multiple sensors, a multi-sensor fusion system fusing various sensory information is needed. In this paper, we propose a Recurrent Fuzzy Inference (RFI) with recurrent inputs and apply it to a multi-sensor fusion system for estimating the state of systems. The membership functions of RFI are expressed by Radial Basis Function (RBF) with insensitive ranges. The shape of the membership functions can be adjusted by a learning algorithm. The learning algorithm is based on the steepest descent method and incremental learning which can add new fuzzy rules. The effectiveness of the multi-sensor fusion system using RFI will be shown through a numerical experiment of moving robot and estimation of surface roughness in grinding process.     

防抱制动系统滑模状态观测和控制系统仿真

该文在考虑不平路面随机激励作用下车辆垂向振动的基础上 ,首先建立了四分之一车辆制动模型 ,而后充分运用滑移模式变结构的分析和设计方法 ,提出了车轮最佳滑移率的滑模实时在线辨识滑模优化算法 ,在对系统可观测性论证的基础上 ,设计了非线性滑模状态观测器 ,给出了单通道防抱制动系统基于滑移率的滑模控制算法 ,通过计算机仿真 ,验证了该控制算法的可行性和有效性 ,为设计具有高鲁棒性的防抱制动系统做了一定的理论探索和仿真工作     

Adaptive motion control of wheeled mobile robot with unknown slippage

As a major representative nonholonomic system, wheeled mobile robot (WMR) is often used to travel across off-road environments that could be unstructured environments. Slippage often occurs when WMR moves in slopes or uneven terrain, and the slippage generates large accumulated position errors in the vehicle, compared with conventional wheeled mobile robots. An estimation of the wheel slip ratio is essential to improve the accuracy of locomotion control. In this paper, we propose an improved adaptive controller to allow WMR to track the desired trajectory under unknown longitudinal slip, where the stabilisation of the closed-loop tracking system is guaranteed by the Lyapunov theory. All system states use neural network online weight tuning algorithms, which ensure small tracking errors and no loss of stability in robot motion with bounded input signals. We demonstrate superior tracking results using the proposed control method in various Matlab simulations.     

GAT: Platform for automatic context-aware mobile services for m-tourism

Despite the recent advances in mobile tourism systems, most of the wayfinding applications have still to deal with some problems: a huge amount of tourist information to manage, guidance for indoor and outdoor environments, and the need of users to have programming knowledge about many mobile phone platforms. In this study, we propose the GAT platform to overcome these problems. In GAT, users are able to generate wayfinding applications for indoor and outdoor environments through a web form without the need for programming skills, assisted by a system of automatic generation and update of points of interest.     

The study on integrating WebQuest with mobile learning for environmental education

This study is to demonstrate the impact of different teaching strategies on the learning performance of environmental education using quantitative methods. Students learned about resource recycling and classification through an instructional website based on the teaching tool of WebQuest. There were 103 sixth-grade students participating in this study and broken down into three groups: traditional instruction, traditional instruction with WebQuest and WebQuest instruction with outdoors. The major contribution of this study is the introduction of WebQuest into the outdoor instruction. The results of this study show that using WebQuest in outdoor instruction influences students’ learning performance positively. Two other interesting results are: (1) when WebQuest was used in real situations, students could acquire more knowledge and experiences, and (2) in the learning activity of the experiment, the students accomplished different learning tasks and expressed their own opinions and perspectives, which could foster their critical thinking skills. On the other hands, the students in outdoor situation could be positive to participate in learning activity; furthermore, they could ponder the learning contents by observing the real context and then they began to classify/categorize the resources. These findings will contribute to the development of teaching and learning for government, schools and teachers; for instance, teachers act as assistants or tutors and provide students with others public network resources, including PDAs, smartphones, tablet personal computers or the Internet, to improve their learning in outdoor learning environments, such as campus, museums or zoos.     

Inventing motivates and prepares student teachers for computer‐based learning

A brief, problem‐oriented phase such as an inventing activity is one potential instructional method for preparing learners not only cognitively but also motivationally for learning. Student teachers often need to overcome motivational barriers in order to use computer‐based learning opportunities. In a preliminary experiment, we found that student teachers who were given paper‐based course material spent more time on follow‐up coursework than teachers who were given a well‐developed computer‐based learning environment (CBLE), leading to higher learning outcomes. Thus, we tested inventing as an instructional method that may help overcome motivational barriers of teachers' use of computer‐supported tools or learning environments in our main experimental study (N = 44). As a computer‐based environment, we used the ‘Assessment of Learning strategies in Learning journals’. The inventing group produced ideas about criteria to evaluate learning strategies based on student cases prior to the learning phase. The control condition read a text containing possible answers to the inventing problem. The inventing activity enhanced motivation prior to the learning phase and assessment skills as assessed by transfer problems. Hence, the inventing activity prepared student teachers to learn from a CBLE in a motivational as well as cognitive way.     

Fuzzy-logic-assisted interacting multiple model (FLAIMM) for mobile robot localization

Improvement of dead reckoning accuracy is essential for robotic localization systems and has been intensively studied. However, existing solutions cannot provide accurate positioning when a robot suffers from changing dynamics such as wheel slip. In this paper, we propose a fuzzy-logic-assisted interacting multiple model (FLAIMM) framework to detect and compensate for wheel slip. Firstly, two different types of extended Kalman filter (EKF) are designed to consider both no-slip and slip dynamics of mobile robots. Then a fuzzy inference system (FIS) model for slip estimation is constructed using an adaptive neuro-fuzzy inference system (ANFIS). The trained model is utilized along with the two EKFs in the FLAIMM framework. The approach is evaluated using real data sets acquired with a robot driving in an indoor environment. The experimental results show that our approach improves position accuracy and works better in slip detection and compensation compared to the conventional multiple model approach.