Abstract:We present a biologically inspired approach for path planning with dynamic obstacle avoidance. Path planning is performed in a condensed configuration space of a robot generated by self-organizing neural networks (SONN). The robot itself and static as well as dynamic obstacles are mapped from the Cartesian task space into the configuration space by precomputed kinematics. The condensed space represents a cognitive map of the environment, which is inspired by place cells and the concept of cognitive maps in mammalian brains. The generation of training data as well as the evaluation were performed on a real industrial robot accompanied by simulations. To evaluate the reactive collision-free online planning within a changing environment, a demonstrator was realized. Then, a comparative study regarding sample-based planners was carried out. So we could show that the robot is able to operate in dynamically changing environments and re-plan its motion trajectories within impressing 0.02 seconds, which proofs the real-time capability of our concept.
What problem does this paper attempt to address?
The problem that this paper attempts to solve is: in a dynamically changing environment, how to achieve fast and collision - free path planning for industrial robots. Specifically, the authors propose a bio - inspired method to deal with the curse of dimensionality problem encountered by traditional path - planning algorithms in high - dimensional configuration space (C - space), especially in the applications of multi - degree - of - freedom (DOF) robots and complex environments.
### Specific Problem Description
1. **Complexity of High - Dimensional Configuration Space**:
- Traditional path - planning algorithms such as Wavefront, Dijkstra, and A* have a huge amount of computation when dealing with spaces of more than 3 dimensions and it is difficult to provide solutions within a reasonable time.
- As the degrees of freedom of the robot increase, the dimension of the configuration space also increases correspondingly, resulting in an exponential increase in the complexity of the search space and the computational cost.
2. **Dynamic Obstacle Avoidance**:
- In a dynamic environment, the robot needs to perceive and avoid moving obstacles in real - time, which places higher requirements on the speed and flexibility of path planning.
- Although sample - based path planners can reduce the computational burden, they still face challenges in high - dimensional spaces and narrow passages and it is difficult to achieve real - time online planning.
3. **Mapping from Cartesian Space to Configuration Space**:
- Mapping obstacles in the task space to the configuration space is a non - trivial task, especially for multi - degree - of - freedom robots, and the required memory increases exponentially with the degrees of freedom.
- Geometric methods and analytical methods are less efficient in dealing with obstacles of complex shapes and cannot meet the requirements of real - time planning.
### Proposed Solution
To overcome the above problems, the authors propose a path - planning method based on self - organizing neural network (SONN), which has the following characteristics:
- **Reducing the Dimension of Configuration Space**: Map the high - dimensional configuration space to a low - dimensional cognitive map through SONN, thereby significantly reducing the complexity of the search space.
- **Real - Time Dynamic Obstacle Avoidance**: Use a bidirectional lookup table to quickly map obstacles in the task space to the configuration space and update the path in real - time.
- **Bio - Inspired Model**: Draw on the concepts of place cells and cognitive maps in the mammalian brain, enabling the robot to perform efficient path planning in complex dynamic environments.
### Experimental Verification
The experimental results show that this method can complete replanning within 0.02 seconds, proving its real - time ability. Compared with traditional sample - based path planners, this method not only performs well in path optimization but also has obvious advantages in execution time and path smoothness.
In conclusion, this paper aims to solve the path - planning problems in high - dimensional configuration spaces, especially to achieve fast, collision - free motion planning in dynamic environments.