Some decades ago, Lotfi A. Zadeh has introduced an unconventional mathematics theory called Fuzzy logic. The concept was rejected by most of his colleagues because mathematics has to do with abstract formalized thinking, and blur the line between zero and one isn't the best idea. Until today, some robotics engineers belief that Fuzzy set theory is a here to stay for modeling robotics problems. In a recent paper an amateur project was described which has a lot of unsolved issues.
The system consists of layers equal natural language terms like “near” and “far”. This is enriched with mathematical equations into a inedible conglomerate. If we belief the paper, then the robot is able to navigate to the goal with the help of a neural network but like most fuzzy motivated projects , the claimed capability can't be proven.
- Motion Planning of Autonomous Mobile Robot Using Recurrent Fuzzy Neural Network Trained by Extended Kalman Filter https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6378056/
- Krzysztof Waszkiewicz: KiKS - The task of following the light https://youtu.be/KVNCydtVuO8