Daniel Hess

Localization of an omnidirectional transport robot using IEEE 802.15.4a ranging and laser range finder

Automated Guided Vehicles (AGVs) are used in warehouses, distribution centers and manufacturing plants in order to automate the internal material flow. Usually AGVs are designed to transport large and heavy transport units such as Euro-pallets or mesh pallets. Just-in-time inventory management and lean production requires small transportation units to enable one-piece-flow. Furthermore short production cycles require a flexible material flow which can not be fulfilled by continuous material handling devices like belt or roll conveyors. A solution to meet these demands are small mobile robots for material transport which can replace conventional conveyor systems or large AGVs. The paper presents localization and tracking of an omnidirectional mobile robot equipped with Mecanum wheels, which was designed to transport Euro-bins in a distribution center or warehouse. Localization is realized by sensor fusion of range measurements obtained from an IEEE 802.15.4a network and laser range finders. The IEEE 802.15.4a network is used for communication as well as for global localization. Laser range finders are used to detect landmarks and to provide accurate positioning for docking maneuvers. The range measurements are fused in a Monte Carlo Particle Filter. The paper develops a new motion model for an omnidirectional robot as well as a sensor model for IEEE 802.15.4a range measurements. The experimental results presented in the paper show the effectiveness of the developed models.

Comparison of trajectory tracking controllers for emergency situations

Over the last years a number of different vehicle controllers has been proposed for tracking planned paths or trajectories. Most of previously published works do not compare their results with other approaches or limit the comparison to a few scenarios. Unfortunately, comparisons with existing controller concepts are very rare and a ranking is hard to establish from the literature. In this work, we rigorously compare inversion-based trajectory tracking controllers by systematically exploring the set of possible solutions when disturbances vary over time and initial states and parameters are uncertain. By using Monte-Carlo simulation, we determine the average performance and by using rapidly exploring random trees, we determine the worst-case performance, which is especially important in emergency situations when avoiding a crash is essential. The tested scenarios and the applied methodologies are documented in detail so that they serve as benchmark problems for other control concepts. The results show that the controller with smaller relative degree performs better with respect to the worst-case deviation computed by rapidly exploring random trees, while conventional simulations of random scenarios would not reveal any difference.

Formal verification of maneuver automata for parameterized motion primitives

An increasing amount of robotic systems is developed for safety-critical scenarios, such as automated cars operating in public road traffic or robots collaborating with humans in flexible manufacturing systems. For this reason, it is important to provide methods that formally verify the safety of robotic systems. This is challenging since robots operate in continuous action spaces in partially unknown environments so that there exists no finite set of scenarios that can be verified before deployment. Verifying the safety during the operation based on the current perception of the environment is often infeasible due to the computational demand of formal verification methods. In this work, we compute sets of behaviors for parameterized motion primitives using reachability analysis, which is used to build a maneuver automaton that connects motion primitives in a safe way. Thus, the computationally expensive task of building a maneuver automaton is performed offline. The proposed analysis method provides the whole set of possible behaviors so that it can be verified whether forbidden state-space regions are avoided during the operation of the robot, to e.g. avoid colliding with obstacles. The method is applied to continuous sets of parameterized motion primitives, making it possible to verify infinitely many motions within the parameter space, which to the best knowledge of the authors has not been published before. The approach is demonstrated for collision avoidance of road vehicles.

Road occupancy prediction of traffic participants

We predict the road occupancy of traffic participants for collision avoidance systems. The occupancy sets are computed for consecutive time intervals and contain all reachable positions of traffic participants in compliance with a proposed dynamic model. Those sets make it possible to check if planned emergency maneuvers are collision-free in all possible future scenarios. However, no algorithm exists for exactly computing the occupancy when the model forbids unrealistic behavior such as leaving road boundaries or largely exceeding speed limits. For this reason, we provide methods to tightly overapproximate occupancy sets to ensure safe emergency maneuvers. We demonstrate the applicability of the approach by numerical examples, which show that the occupancy computation is not only efficient, but also tight enough to trigger emergency maneuvers almost at the last possible point in time.

Contingency planning for automated vehicles

Automated driving is a safety critical process, which requires complex decision making. In order to validate driving decisions, it is possible to maintain at all times a contingency maneuver, which transfers the vehicle to a safe standstill, if other decision making processes fail. In this paper we present a motion planner, which computes contingency maneuvers for an automated vehicle in a 0.1[s] time frame. A discrete set of motion primitives is assembled in a heuristic best-first search. In order to speed up the search, an obstacle sensitive heuristic is applied, which maintains properties of bounded sub-optimality and completeness. A run-time comparison with and without the obstacle sensitive heuristic is presented on two exemplary collision avoidance scenarios.

Safe@home - A wireless assistance system with integrated IEEE 802.15.4a localisation technology

The wireless assistance system safe@home builds a wireless sensor network which enables the localisation of a person wearing a tag and to collect the persons vital data. Furthermore it provides the opportunity to connect to home automation elements like EnOcean1 sensors and actuators. The goal to this project is to provide a system which can easily be installed into existing homes and therefore is low-priced, to allow elder people to live safely and independently in their own homes for as long as possible.

Double-lane change optimization for a stochastic vehicle model subject to collision probability constraints

This paper presents an approach to autonomous passenger vehicle path planning, which accounts for the probability of collision arising from noise affected motion and imprecise future observations of the vehicle state. The probabilities of future system states are approximated by Gaussian distributions and the mean and covariance trajectories are predicted for the closed-loop system consisting of vehicle and path following controller. Linear approximations of vehicle model and controller at multiple points of the state mean trajectory relate the future covariance to the path under optimization. The situation dependent constraints on safety distance to obstacles, imposed by the dependent closed-loop covariance and a limit to the probability of collision, lead to more accurate results than a static margin of error or a distance maximization objective as currently employed by other state of the art path planners. The reference path optimization under collision probability constraints is reduced to a deterministic and static optimization problem and formulated as a nonlinear program. Numerical results for the optimization of two example double-lane change maneuvers illustrate the benefits of the approach.

Should collision avoidance systems use yaw stabilization

Due to historical reasons or system development aspects, many high-level control tasks in vehicles are performed by underlying low-level controllers. This separation of concerns provides reliable systems, but potentially degrades the performance compared to centralized control. Performance losses are acceptable for most control tasks, but for collision avoidance systems one should not compromise on safety. We investigate the performance loss for collision avoidance systems when an underlying yaw stabilization controller is used, which can be found in many modern vehicles under various product names, such as electronic stability control (ESC). Since electronic stability control differs from vehicle to vehicle, we use an idealized controller that performs better than or equally well as an actual realization. It is shown that central control concepts bypassing the yaw stabilization perform better than a hypothetical controller embedded with the idealized yaw stabilization. We also provide a measure for the performance loss, which should support the decision for or against the use of yaw stabilization in collision avoidance systems.