Path Planning
Path planning facilitates global and local planning algorithms in order to find a feasible route to target coordinates, using both real-time sensory information and environment maps. It is the basic requirement for directed motion of a mobile robot to its target within the context of autonomous navigation. Global planning focuses primarily on devising a rough path towards a target, based on a map of the environment. Local planning takes dynamic obstacles and collision avoidance into account.
INTRODUCTION
Robust Trajectory Generation for dynamic and unpredictable Environments
Ongoing development enhances the capabilities of dynamic obstacle avoidance and incorporates individual strategies for certain spatial scenarios, e.g. for avoiding local minima or predicting human behavior.
Gestalt employs a large spectrum of path planning algorithms. Additionally, we specialize in individual solutions for scenarios in which established path planning algorithms have to be extended in order to meet particular requirements of the environment.
Collision avoidance by sampling-based path planning and subsequent smoothing
TECHNOLOGICAL OUTLINE
The right Tool for the Job
There is a huge variety of motion scenarios, each posing its own set of intricate constraints to path planning algorithms. Hence, the portfolio of established approaches to tackle them is large and has to put up with new challenges posed by advancing technologies like drone control and human-robot cooperation in industrial manufacturing.
Motion planning in deterministic assembly-line scenarios is mostly optimized for speed and robot endurance, the path optimization can be complex because the planning happens only once. In human-robot interaction, the main concerns are safety and predictability: Fast dynamic replanning must be possible in order to cope with sudden obstacles and events. Vehicular robots are best investigated in Cartesian space, manipulators with bulky tools in collision-prone scenarios are better handled in configuration space. For simple environments, a complete map of the surroundings can be explored (potential fields, A*, D*), in more complex situations, sampling-based algorithms (probabilistic roadmap, tree-based planners) are better suited.
However, although numerous approaches are already established for the different circumstances, each individual application poses special challenges and in most cases requires distinct creative solutions, tailored to its specific requirements.
APPLICATIONS & USE CASES
Industrial Manipulators – Path Generation and Optimization
For industrial partners we develop, implement and integrate tailored path and motion planning algorithms for industrial manipulators. The robot trajectory is to be optimized with respect to different criteria, e.g. cycle times, work spaces, dynamics as well as process and technology parameters. Consequently, each field of application in robotics has its own requirements towards path planning. We typically consider extended demands for thermal cutting, welding, milling and more.
With respect to welding for instance, the sequence in which the seams are passed has to be tweaked, the overall process time is to be minimized, the wear on the robot has to be kept small. The workspace of both the robot and the tool have to be considered and special requirements of the welding process itself have to be taken into account.
Iterative sequence optimization for welding tasks
Mobile service robotic navigation – Safe and context-aware
We are working with ROS-based mobile platform solutions with navigation and obstacle avoidance. State of the art path planning algorithms facilitate real-time reaction to external influences and robust replanning. Furthermore, we customize ROS-based path planning and navigation in order to optimize efforts, consider dynamic obstacles, adapt algorithms to specific environmental constraints and to be context-aware.
Hence, we integrate robotic service assistants seamlessly into naturally human environments like museums, hospitals, event venues and store houses, being perceived as helpful and non-obtrusive.
On-the-fly path planning and replanning depending on recognized obstacles
CONCLUSION
Path planning remains one of the fundamental aspects of robot control. Collision-free, reachable, smooth, safe, predictable and reactive path planning is vital for integrating robots into both industry and society. GESTALT utilizes a broad spectrum of established motion planning algorithms and adapts them to the special circumstances of the individual customer. Don't hesitate to contact us and discuss your use case via info@gestalt-robotics.com or give us a call at +49 30 616 515 60 - we would love to hear from you.