A discrete-time stable controller for an omni-directional mobile robot based on an approximated model

An adaptive controller based on multi-input fuzzy rules emulated networks (MIFRENs) is introduced for omni-directional mobile robot systems in the discrete-time domain without any kinematic or dynamic models. An approximated model for unknown systems is developed by using two MIFRENs with an online learning algorithm in addition to the stability analysis. The main theorem in this model is proposed to guarantee closed-loop performance and system robustness for all adjustable parameters inside MIFRENs. The system is validated by an experimental setup with a FESTO omni-directional mobile robot called Robotino^®. The proposed algorithm is shown to have superior performance compared to that of an algorithm that uses only an embedded controller. The advantage of the MIFREN initial setting is verified comparing its results with those of a controller that is based on neural networks.     

Introducing Synaptic Delays in the NEAT Algorithm to Improve Modelling in Cognitive Robotics

This paper describes and tests an approach to improve the temporal processing capabilities of the neuroevolution of augmenting topologies (NEAT) algorithm. This algorithm is quite popular within the robotics community for the production of trained neural networks without having to determine a priori their size and topology. The main drawback of the traditional NEAT algorithm is that, even though it can implement recurrent synaptic connections, which allow it to perform some time related processing tasks, its capabilities are rather limited, especially when dealing with precise time dependent phenomena. NEAT’s ability to capture the underlying dynamics that correspond to complex time series still has a lot of room for improvement. To address this issue, the paper describes a new implementation of the NEAT algorithm that is able to generate artificial neural networks (ANNs) with trainable time delayed synapses in addition to its previous capacities. We show that this approach, called \(\uptau \)-NEAT improves the behavior of the neural networks obtained when dealing with complex time related processes. Several examples are presented, both dealing with the generation of ANNs that are able to produce complex theoretical signals such as chaotic signals or real data series, as in the case of the monthly number of international airline passengers or monthly \(\hbox {CO}_{2}\) concentrations. In these examples, \(\uptau \)-NEAT clearly improves over the traditional NEAT algorithm in these tasks. A final example of the integration of this approach within a robot cognitive mechanism is also presented, showing the clear improvements it could provide in the modeling required for many cognitive processes.     

Real time implementation of CTRNN and BPTT algorithm to learn on-line biped robot balance: Experiments on the standing posture

This paper describes experimental results regarding the real time implementation of continuous time recurrent neural networks (CTRNN) and the dynamic back-propagation through time (BPTT) algorithm for the on-line learning control laws. Experiments are carried out to control the balance of a biped robot prototype in its standing posture. The neural controller is trained to compensate for external perturbations by controlling the torso’s joint motions. Algorithms are embedded in the real time electronic unit of the robot. On-line learning implementations are presented in detail. The results on learning behavior and control performance demonstrate the strength and the efficiency of the proposed approach.     

Self-calibration of environmental camera for mobile robot navigation

An environmental camera is a camera embedded in a working environment to provide vision guidance to a mobile robot. In the setup of such robot systems, the relative position and orientation between the mobile robot and the environmental camera are parameters that must unavoidably be calibrated. Traditionally, because the configuration of the robot system is task-driven, these kinds of external parameters of the camera are measured separately and should be measured each time a task is to be performed. In this paper, a method is proposed for the robot system in which calibration of the environmental camera is rendered by the robot system itself on the spot after a system is set up. Specific kinds of motion patterns of the mobile robot, which are called test motions, have been explored for calibration. The calibration approach is based upon executing certain selected test motions on the mobile robot and then using the camera to observe the robot. According to a comparison of odometry and sensing data, the external parameters of the camera can be calibrated. Furthermore, an evaluation index (virtual sensing error) has been developed for the selection and optimization of test motions to obtain good calibration performance. All the test motion patterns are computed offline in advance and saved in a database, which greatly shorten the calibration time. Simulations and experiments verified the effectiveness of the proposed method.     

Identification of a golf swing robot using soft computing approach

Golf swing robots have been recently developed in an attempt to simulate the ultra high-speed swing motions of golfers. Accurate identification of a golf swing robot is an important and challenging research topic, which has been regarded as a fundamental basis in the motion analysis and control of the robots. But there have been few studies conducted on the golf swing robot identification, and comparative analyses using different kinds of soft computing methodologies have not been found in the literature. This paper investigates the identification of a golf swing robot based on four kinds of soft computing methods, including feedforward neural networks (FFNN), dynamic recurrent neural networks (DRNN), fuzzy neural networks (FNN) and dynamic recurrent fuzzy neural networks (DRFNN). The performance comparison is evaluated based on three sets of swing trajectory data with different boundary conditions. The sensitivity of the results to the changes in system structure and learning rate is also investigated. The results suggest that both FNN and DRFNN can be used as a soft computing method to identify a golf robot more accurately than FFNN and DRNN, which can be used in the motion control of the robot.     

Dynamical neural networks for planning and low-level robot control

We use dynamical neural networks based on the neural field formalism for the control of a mobile robot. The robot navigates in an open environment and is able to plan a path for reaching a particular goal. We will describe how this dynamical approach may be used by a high level system (planning) for controlling a low level behavior (speed of the robot). We give also results about the control of the orientation of a camera and a robot body.     

基于模糊神经网络的移动机器人沿墙导航控制

移动机器人沿墙导航控制包含了追踪和避障两种情况,是移动机器人研究中的常见问题。它是指机器人在一定方向上沿墙运动,或者更一般意义上的沿着物体轮廓运动,并与墙保持一定距离。移动机器人利用声纳采集机器人与墙体的距离和角度信息,通过模糊神经网络将输入数据进行融合,从而判断移动机器人的位姿信息,输出左右轮速度控制其动作。实验证明此方法可以有效地保证移动机器人在安全距离内沿墙体运动。对比采用模糊神经网络前后的实验,采用后的移动机器人沿墙导航控制轨迹优于采用前,均方误差大大减小。     

Modelling the Environment of a Mobile Robot with the Embedded Flow State Machine

A type of topological approach to mobile robot navigation is discussed and experimentally evaluated. The environment as experienced by a moving robot is treated as a dynamical system. Simple types of reactive behavior are supplemented with eventual decisions to switch between them. When switching criteria are defined, the system may be described in the form similar to a finite state machine. Since it is embedded in the environment and dependent on the sensory flow of the robot, we introduce the term “Embedded flow state machine” (EFSM). We implemented it with a recurrent neural network, trained on a sequence of sensory contents and actions. One of the main virtues of this approach is that no explicit localization is required, since the recurrent neural network holds the state implicitly. The EFSM is applicable to multi-step prediction of sensory information and the travelled distances between decision points, given a sequence of decisions at decision points. Thus, the optimal path to a specified goal can be sought. One of the main issues is, for how many steps ahead the prediction is reliable enough. In other words, is it feasible to perform environment modelling and path planning in this manner? The approach is tested on a miniature mobile robot, equipped with proximity sensors and a color video camera. Decision ‘points,’ where deviations from the wall-following behavior are allowed, are based on color object recognition. In the case of an experimental environment of medium complexity, this approach was successful.     

A neural network-based approach for an assembly cell control

In several robotics applications, the robot must interact with the workspace, and thus its motion is constrained by the task. In this case, pure position control will be ineffective since forces appearing during the contacts must also be controlled. However, simultaneous position and force control called hybrid control is then required. Moreover, the nonlinear plant dynamics, the complexity of the dynamic parameters determination and computation constraints makes more difficult the synthesis of control laws. In order to satisfy all these constraints, an effective hybrid force/position approach based on artificial neural networks for multi-inputs/multi-outputs systems is proposed. This approach realizes, simultaneously, an identification and control of systems, and it is implemented according to two phases: At first, a neural observer is trained off-line on the basis of the data acquired during contact motion, in order to realize a smooth transition from free to contact motion. Then, an online learning of the neural controller is implemented using neural observer parameters so that the closed-loop system maintains a good performance and compensates for uncertain/unknown dynamics of the robot and the environment. A typical example on which we shall focus is an assembly task. Experimental results on a C5 links parallel robot demonstrate that the robot's skill improves effectively and the force control performances are satisfactory, even if the dynamics of the robot and the environment change.     

Real-Time Dynamic Visual Tracking Using PSD Sensors and Extended Trapezoidal Motion Planning

A real-time visual servo tracking system for an industrial robot has been implemented using PSD (Position Sensitive Detector) cameras, neural networks, and an extended trapezoidal motion planning method. PSD and directly transduces the light's projected position on its sensor plane into an analog current and lends itself to fast real-time tracking. A neural network, after proper training, transforms the PSD sensor reading into a 3D position of the target, which is then input to an extended trapezoidal motion planning algorithm. This algorithm implements a continuous motion update strategy in response to an ever-changing sensor information from the moving target, while greatly reducing the tracking delay. This planning method is found to be very useful for sensor-based control such as moving target tracking or weld-seam tracking in which the robot needs to change its motion in real time in response to incoming sensor information. Further, for real-time usage of the neural net, a new architecture called LANN (Locally Activated Neural Network) has been developed based on the concept of CMAC input partitioning and local learning. Experimental evidence shows that an industrial robot can smoothly track a moving target of unknown motion with speeds of up to 1 m/s and with oscillation frequency up to 5 Hz.     

Autonomous mobile robot dynamic motion planning using hybrid fuzzy potential field

A new fuzzy-based potential field method is presented in this paper for autonomous mobile robot motion planning with dynamic environments including static or moving target and obstacles. Two fuzzy Mamdani and TSK models have been used to develop the total attractive and repulsive forces acting on the mobile robot. The attractive and repulsive forces were estimated using four inputs representing the relative position and velocity between the target and the robot in the x and y directions, in one hand, and between the obstacle and the robot, on the other hand. The proposed fuzzy potential field motion planning was investigated based on several conducted MATLAB simulation scenarios for robot motion planning within realistic dynamic environments. As it was noticed from these simulations that the proposed approach was able to provide the robot with collision-free path to softly land on the moving target and solve the local minimum problem within any stationary or dynamic environment compared to other potential field-based approaches.     

Bidirectional segmented-memory recurrent neural network for protein secondary structure prediction

The formation of protein secondary structure especially the regions of β-sheets involves long-range interactions between amino acids. We propose a novel recurrent neural network architecture called segmented-memory recurrent neural network (SMRNN) and present experimental results showing that SMRNN outperforms conventional recurrent neural networks on long-term dependency problems. In order to capture long-term dependencies in protein sequences for secondary structure prediction, we develop a predictor based on bidirectional segmented-memory recurrent neural network (BSMRNN), which is a noncausal generalization of SMRNN. In comparison with the existing predictor based on bidirectional recurrent neural network (BRNN), the BSMRNN predictor can improve prediction performance especially the recognition accuracy of β-sheets.     

Simultaneous localization of multiple unknown CSMA-based wireless sensor network nodes using a mobile robot with a directional antenna

We use a single mobile robot equipped with a directional antenna to simultaneously localize unknown carrier sensing multiple access (CSMA)-based wireless sensor network nodes. We assume the robot can only sense radio transmissions at the physical layer. The robot does not know network configuration such as size and protocol. We formulate this new localization problem and propose a particle filter-based localization approach. We combine a CSMA model and a directional antenna model using multiple particle filters. The CSMA model provides network configuration data while the directional antenna model provides inputs for particle filters to update. Based on the particle distribution, we propose a robot motion planning algorithm that assists the robot to efficiently traverse the field to search radio source. The final localization scheme consists of two algorithms: a sensing algorithms that runs in O(n) time for n particles and a motion planning algorithm that runs in O(nl) time for l radio sources. We have implemented the algorithm, and the results show that the algorithms are capable of localizing unknown networked radio sources effectively and robustly.     

Method for tuning the motion planning parameters of a mobile robot

This paper presents a method for configuring the motion planning system of an omniwheeled mobile robot with a differential drive. A simulation program that models the horizontal movement of the robot is described. This simulation program is used to select the optimal parameters for the differential drive control algorithm. Then, the motion planning system is tested on a real robot, which is called RB-2, to adjust the parameters selected. This approach allows the control algorithm to be tuned efficiently and effectively, minimizing the number of its test runs on the physical robot.     

Adaptive Behaviors in Autonomous Navigation with Collision Avoidance and Bounded Velocity of an Omnidirectional Mobile Robot

Integration of Control Theory and Genetic Programming paradigm toward development a family of controllers is addressed in this paper. These controllers are applied for autonomous navigation with collision avoidance and bounded velocity of an omnidirectional mobile robot. We introduce the concepts of natural and adaptive behaviors to relate each control objective with a desired behavior for the mobile robot. Natural behaviors lead the system to fulfill a task inherently. In this work, the motion of the mobile robot to achieve desired position, ensured by applying a Control-Theory-based controller, defines the natural behavior. The adaptive behavior, learned through Genetic-Programming, fits the robot motion in order to avoid collision with an obstacle while fulfilling velocity constraints. Hence, the behavior of the mobile robot is the addition of the natural and the adaptive behaviors. Our proposed methodology achieves the discovery of 9402 behaviors without collisions where asymptotic convergence to desired goal position is demonstrated by Lyapunov stability theory. Effectiveness of proposed framework is illustrated through a comparison between experiments and numerical simulations for a real mobile robot.     

Path planning in GPS-denied environments via collective intelligence of distributed sensor networks

This paper proposes a framework for reactive goal-directed navigation without global positioning facilities in unknown dynamic environments. A mobile sensor network is used for localising regions of interest for path planning of an autonomous mobile robot. The underlying theory is an extension of a generalised gossip algorithm that has been recently developed in a language-measure-theoretic setting. The algorithm has been used to propagate local decisions of target detection over a mobile sensor network and thus, it generates a belief map for the detected target over the network. In this setting, an autonomous mobile robot may communicate only with a few mobile sensing nodes in its own neighbourhood and localise itself relative to the communicating nodes with bounded uncertainties. The robot makes use of the knowledge based on the belief of the mobile sensors to generate a sequence of way-points, leading to a possible goal. The estimated way-points are used by a sampling-based motion planning algorithm to generate feasible trajectories for the robot. The proposed concept has been validated by numerical simulation on a mobile sensor network test-bed and a Dubin’s car-like robot.     

Experience-based imitation using RNNPB

Robot imitation is a useful and promising alternative to robot programming. Robot imitation involves two crucial issues. The first is how a robot can imitate a human whose physical structure and properties differ greatly from its own. The second is how the robot can generate various motions from finite programmable patterns (generalization). This paper describes a novel approach to robot imitation based on its own physical experiences. We considered the target task of moving an object on a table. For imitation, we focused on an active sensing process in which the robot acquires the relation between the object's motion and its own arm motion. For generalization, we applied the RNNPB (recurrent neural network with parametric bias) model to enable recognition/generation of imitation motions. The robot associates the arm motion which reproduces the observed object's motion presented by a human operator. Experimental results proved the generalization capability of our method, which enables the robot to imitate not only motion it has experienced, but also unknown motion through nonlinear combination of the experienced motions.     

Navigation behavior based on self-organized spatial representation in hierarchical recurrent neural network

ABSTRACT

A cognitive map is an internal model of the external world and contains the spatial representation of the surrounding environment. The existence of the cognitive map was first identified in rats; rats can navigate to their desired destination using cognitive maps while dealing with environmental uncertainty. We performed a mobile robot navigation experiment where obstacles were randomly placed using hierarchical recurrent neural network (HRNN) with multiple timescales. The HRNN was trained to navigate the mobile robot to the destination indicated by a snapshot image. After the training, the HRNN was able to successfully avoid the obstacles and navigate to the destination from any location in the environment. Analysis of the internal states of the HRNN showed that the module with fast timescale handles obstacle avoidance and the one with slow timescale has spatial representation corresponding to the spatial position of the destination. Moreover, in the experiment wherein the novel path appeared, the trained HRNN performed shortcut behavior. The shortcut behavior shows that the HRNN performed navigation using the self-organized spatial representation in the slow recurrent neural network. This indicates that training of goal-oriented navigation, i.e. the navigation motivated by a snapshot image of the destination results in the self-organization of cognitive map-like representation.     

Existence, learning, and replication of periodic motions inrecurrent neural networks

A class of recurrent neural networks is shown to possess a stable limit cycle. A gradient type algorithm is used to modify the parameters of the network so that it learns and replicates autonomously a time varying periodic signal. The results are applied to controlling the repetitive motion of a two-link robot manipulator.     

Improving Robustness of Mobile Robots Using Model-based Reasoning

Retaining functionality of a mobile robot in the presence of faults is of particular interest in autonomous robotics. From our experiences in robotics we know that hardware is one of the weak points in mobile robots. In this paper we present the foundations of a system that automatically monitors the driving device of a mobile robot. In case of a detected fault, e.g., a broken motor, the system automatically reconfigures the robot in order to still allow to reach a certain position. The described system is based on a generalized model of the motion hardware. High-level control like path-planner only to change its behavior in case of a serious damage. The high-level control system remains the same. In the paper we present the model and the foundations of the diagnosis and reconfiguration system. This research has been funded in part by the Austrian Science Fund (FWF) under grant P17963-N04. Authors are listed in alphabetic order.