Amos Albert

Experiments in vision-guided biped walking

Goal-oriented vision-guided biped locomotion requires a high degree of coordination between perception and walking. How to establish this coordination remains a fundamental and rarely studied problem in legged robotics. Some of our investigations into this field are outlined in this article by presenting recent results in vision-guided biped locomotion. The guidance and control approaches developed are experimentally validated on the biped robot BARt-UH. It is shown how perception techniques are employed in closed-loop for step sequence adaptation and locomotion control of a walking machine.

Analytic Path Planning Algorithms for Bipedal Robots without a Trunk

This paper presents new path planning algorithms for bipedal robots without a trunk. The proposed methods can be regarded as an extension to the concept of the inverted pendulum mode (IPM). Further, they fill the gap between the IPM and concepts, based on general dynamic modeling. The swing leg motion is adjusted almost arbitrarily and the torso motion is calculated analytically in order to ensure stability. The proposed methods show a higher gait stability compared to literature approaches. This improvement is verified by measurement of the foot reaction forces during the walk of the existing biped BARt-UH and simulations.

Wireless node localization based on RSSI using a rotating antenna on a mobile robot

Determining the physical location of radio nodes is crucial in a wireless sensor network in order to geo-reference the sensor information. This paper presents a localization technique using a mobile radio node attached to a robot. Its position is given by a real time kinematic (RTK) GPS. We propose a new method to obtain the relative bearing between the mobile and the static node using commodity radio hardware. The method exploits the anisotropy of the antenna gain and is not only more precise than the mapping of the RSSI to distance only, it also allows to monitor the integrity of the measurement. Experimental results in an outdoor environment show a standard deviation of 5. for the angle of arrival. The applicability of the approach for the localization of static radio nodes is demonstated by means of a particle filter. The mean error of the position for experimental data is approximately 13 cm.

Detection of stair dimensions for the path planning of a bipedal robot

This paper deals with the detection of the characteristics of stairs, i.e. the number of steps, the step height and the step width for online path planning of a bipedal robot. For the construction of a multi-purpose, mobile platform for service robot applications with special respect to the human environment, a biped is more advantageous than a wheel based robot. In the framework of our studies, the bipedal robot BARt-UH has been built and walking as well as the climbing of stairs have been realized. The environment of the robot is assumed to be structured, consisting of flat surfaces and stairs, but not known in advance. Therefore, a state transition algorithm for intelligent path planning of the robot is suggested. Further, a stereo vision module with a line laser is considered, in order to detect the stair dimensions. The necessary information is extracted from the projection of the line laser onto the stairs. To achieve this, three algorithms were implemented and a comparative study was carried out.

Vision-based high-speed manipulation for robotic ultra-precise weed control

In this system paper we present a novel approach for autonomous, mobile manipulation for agricultural robots. Our target application is mechanical weed control which for example is needed in organic farming. Today, this task is often performed by field workers, whose availability is declining and the quality of their work differs greatly. In addition, manual weeding is a very tedious task with adverse health effects. Hence, mechanical weed control is considered as a prime example for automation. One of the challenges in order to perform the task efficiently is the need for fast plant detection and precise treatment of single plants with a mobile manipulator. The unstructured environment and rough field conditions raise additional challenges. We introduce a first system concept and control design known to handle these issues. In particular, high speed image processing and visual servoing are applied to precisely position a specially designed weeding tool. To evaluate the performance of the mobile manipulator we run experiments with a mobile test vehicle in field-like conditions and demonstrate that our system can treat single plants in less than 1s as it is required to be competitive.

Radiation pattern correlation for mobile robot localization in low power wireless networks

We present a new method for localization using received signal strength indicator (RSSI) in ordinary wireless communication networks such as specified by IEEE 802.15.4. The method exploits the anisotropy of the antenna gain to determine the bearing of the robot relative to reference radio nodes. This method is not only more precise than the mapping of the RSSI to distance only, it also allows to estimate the orientation of the robot and to monitor the integrity of the measurement. The integrity measure is also incorporated into the RSSI to distance mapping and a thorough error analysis is presented. The paper describes the localization concept and presents experimental results for mobile robot localization in an outdoor environment. The achieved accuracy is significantly increased compared to previously developed RSSI based localization methods.

Extrem kompaktes SoC-Konzept eines Gleichgewichtsorganes für einen Laufroboter

Zur geeigneten Bahnplanung und Regelung eines zweibeinigen Roboters benotigt man einen Sensor, der — ahnlich dem Gleichgewichtsorgan des Menschen — die Lage und Orientierung im Raum messen kann. Dazu erfolgte der Aufbau einer Messapparatur, die im Wesentlichen aus drei Beschleunigungs-, drei Drehraten- und drei Magnetfeldsensoren besteht. Das Problem der Sensorausrichtung fur eine raumliche Messung lies sich auf einfache Weise durch einen geschickten Aufbau in Wurfelform losen. Dieser Artikel beschreibt den mechanischen und elektronischen Aufbau des Gleichgewichtsorgans, die zum Einsatz kommenden Sensoren sowie deren Kalibrierung. Die Bestimmung der gewunschten Lageinformation erfolgt durch Beobachter, die eine Kompensation von Sensordrift und Einbauungenauigkeiten ermoglichen. Die abschliesenden Bemerkungen des Artikels widmen sich der leistungsfahigen Einheit aus dem Mikrocontroller MPC555 und dem Echtzeitbetriebssystem RTOS-UH in seiner Spezialversion SoC (System on the Chip). Erst diese Einheit erlaubte die kompakte Bauweise des Gleichgewichtsorgans.

Approach Towards Robotic Mechanical Weed Regulation in Organic Farming

This paper deals with robotic mechanical weed regulation in organic farming, particularly for carrot cultivation. For that purpose the autonomous agriculture robot ‘BoniRob’ is utilized which is the result of a predecessor project and which allows an ‘App’-concept with changing sensor/actuators arrangements to cope with different use cases.The perception and navigation system is based on semantic localization. It enables adaptation to different environmental conditions encountered in typical tasks. The paper illustrates how this system will now be employed for the task of mechanical weed control. Additionally, the system architecture is described including means to increase robustness and preventing undesirable system conditions. In order to ensure a robust task fulfillment in weed control a shared autonomy approach is proposed which combines an efficient collaboration of the autonomous robot with human interaction via immersion technologies. Further, the paper sketches the ongoing development of the weed manipulator which needs to operate in harsh environments and which is faced with challenging requirements from speed and accuracy perspective. A parallel-kinematic structure enhanced by computer vision and visual servoing is proposed to cope with the requirements. Finally, the paper presents our first results regarding the selection of the actuator principle.

Simultaneous mobile robot and radio node localization in wireless networks

Determining the physical location is a fundamental challenge in location based services and service robotics. This paper presents an approach to simultaneously determine the pose of a mobile robot and the positions of static wireless nodes based on a new technique to compute the angle of arrival of radio signals. For the overall localization process an Extended Kalman Filter is employed. For the initialization of the nodes positions a Particle Filter is used, to overcome the nonlinearity problem. Experiments show a mean accuracy for the position of the robot of 39mm for an indoor office environment and 80mm for an outdoor landscape environment.

Echtzeitsystem für einen zweibeinigen Roboter

Die Einsatzgebiete von Servicerobotern unterliegen oft stetigen Veranderungen oder sind a priori ganzlich unbekannt. Diese Unsicherheiten erfordern eine Roboterstruktur, die eine autonome Navigation erlaubt. Zu berucksichtigende Aspekte sind hierbei der mechanische Aufbau, die Leistungsversorgung, die Sensorik sowie die Rechnerplattform. Dieser Artikel stellt das fur den am Institut fur Regelungstechnik entwickelten zweibeinigen Roboter BARt-UH verwendete Software- und Hardwarekonzept vor. Die Synthese von Software- und Hardware ist von grundlegender Bedeutung, um eine hohe Leistungsfahigkeit und Zuverlassigkeit zu gewahrleisten. Fur die Navigation des Roboters kommt der Mikrocontroller MPC555 aus der PowerPC-Familie und das Multitasking-Echtzeit-Betriebssystem RTOS-UH zum Einsatz.