Auto-lands an object to a specific point on the surface. See the MecanicaScience website for a live demo.
This project aims to let an AI land a vessel on a simulated 2D terrain with a simulated atmosphere, with wind and atmospheric disturbances, while targeting a specific point and minimizing vessel's fuel.
- The awesome book "The Nature of Code" by Daniel Shiffman "The Nature Of Code"
- The NeuroEvolution of Augmented Topologies fondating paper "NEAT"
- The Realtime NeuroEvolution of Augmented Topologies fondating paper (built for videos games) "rtNEAT"
- G-Fold algorithm : how to land with a precision < 100 m on an other planet. G-Fold auto-computes the shortest way in terms of fuel to land on a specific target, based on the current position and velocity of the vessel (witch are unknown due to wind, and atmospheric instabilities) "G-FOLD: A Real-Time Implementable Fuel Optimal Large Divert Guidance Algorithm for Planetary PinpointLanding"
- Landing fuel optimization for engines : "Meditch, J. (1964). On the problem of optimal thrust programming for a lunar soft landing. IEEE Transactions on Automatic Control, 9(4), 477–484. doi:10.1109/tac.1964.1105758"
- Convex programming algorithm for the numerical solution of the minimum fuel powered descent guidance problem associated with Mars pinpoint landing : "Convex Programming Approach to Powered Descent Guidance for Mars Landing"
- Landing cone optimisation : "Lossless Convexification of Nonconvex ControlBound and Pointing Constraints of the SoftLanding Optimal Control Problem"
- Rocket landing game and Deep-Learning algorithm in HTML5 "openai/gym/gym/envs/box2d/lunar_lander.py"
- G-Fold algorithm python implementation "jonnyhyman / G-FOLD-Python"
- Main physics Simulation
- Terrain Generation
- Atmosphere simulation
- Wind simulation
- Collision detection
- Drag forces simulation
- Auto-lander (2 methods : G-Fold and Deep Reinforcement Learning)
- Implementation in KSP