Pinned Repositories
7_of_Hearts
This project will implement a gym environment for the cards game 7 of Hearts using Pygame library and Openai Gym library. Then many RL algorithms can be implemented on this environment.
BehaviorPlanningRL
RL based Behaviour Planning for Autonomous Vehicles using Hierarchal Reinforcement Learning with auto-curriculum generator
BoundingVolumeHierarchyBasedCollisionDetection
Bounding Volume Hierarchy is used to represent the objects in a 3D euclidean space. Collision Detection between this complex detailed structures is implemented in this project.
Centralized_Motion_Planner_Project
Localized motion planning on-board an Autonomous Vehicle (AV) using SLAM (Simultaneous Localization and Mapping) has been implemented. This individual motion planning of the route by AVs does not account for other AVs' routes, urgency, other such parameters and can lead to deadlock situations. A centralized computation unit that provides the registered AVs with a optimal motion plan is proposed by this project. A central planner will find a optimal motion plan for all the AVs in its control area. This central planner will consider the fuel constraint, priorities, speed constraint, etc while planning. This project proposes the use of Particle Swarm Optimization algorithm for meta-heuristic centralized decoupled motion planning. A classic decoupled centralized motion planning algorithm will be implemented for baseline comparison. Advanced algorithm like Grey Wolf Optimization will be studied and explored in the project.
CreepyCrawler-Quadruped-Robot
Quadruped robots have been of interest to the robotics community as they have high maneuverability and they can be designed to be very robust to the environment they are placed in. The use of quadrupeds as service robots is being realized by the community. Understanding the dynamics of a quadruped robot is a complex task as there are many factors that need to be taken into consideration while designing the robot. This project intends to develop an eight degree of freedom (DOF) sprawling-type quadruped robot. Control of high DOF robots can be highly challenging and computationally heavy. This project intends to work on bringing down the complexity of the robot control by decoupling the dynamics of the robot into four separate dynamic systems, i.e one for each leg, allowing us to achieve a decentralized control of the robot. Gazebo simulation of the robot is achieved with a few gait implementations using a decentralized control strategy.
DS595-RL-Projects
Face-Recognition-and-Pose-Classifiaction
PandaPickAndPlace
ROS2Planner
ROS2, Pygame, Discreet Planning Algorithms and Sampling Algorithms.
SLAM_Bot
This project intends to develop an omnidirectional robot car with stereo vision camera and solid state lidar sensor. It will further explore SLAM, RL based control, and autonomous driving stack. The first goal is to add ROS2 based control to the raspberry pi controlled robot car.
RushiPDeshmukh's Repositories
RushiPDeshmukh/Centralized_Motion_Planner_Project
Localized motion planning on-board an Autonomous Vehicle (AV) using SLAM (Simultaneous Localization and Mapping) has been implemented. This individual motion planning of the route by AVs does not account for other AVs' routes, urgency, other such parameters and can lead to deadlock situations. A centralized computation unit that provides the registered AVs with a optimal motion plan is proposed by this project. A central planner will find a optimal motion plan for all the AVs in its control area. This central planner will consider the fuel constraint, priorities, speed constraint, etc while planning. This project proposes the use of Particle Swarm Optimization algorithm for meta-heuristic centralized decoupled motion planning. A classic decoupled centralized motion planning algorithm will be implemented for baseline comparison. Advanced algorithm like Grey Wolf Optimization will be studied and explored in the project.
RushiPDeshmukh/SLAM_Bot
This project intends to develop an omnidirectional robot car with stereo vision camera and solid state lidar sensor. It will further explore SLAM, RL based control, and autonomous driving stack. The first goal is to add ROS2 based control to the raspberry pi controlled robot car.
RushiPDeshmukh/BehaviorPlanningRL
RL based Behaviour Planning for Autonomous Vehicles using Hierarchal Reinforcement Learning with auto-curriculum generator
RushiPDeshmukh/BoundingVolumeHierarchyBasedCollisionDetection
Bounding Volume Hierarchy is used to represent the objects in a 3D euclidean space. Collision Detection between this complex detailed structures is implemented in this project.
RushiPDeshmukh/7_of_Hearts
This project will implement a gym environment for the cards game 7 of Hearts using Pygame library and Openai Gym library. Then many RL algorithms can be implemented on this environment.
RushiPDeshmukh/CreepyCrawler-Quadruped-Robot
Quadruped robots have been of interest to the robotics community as they have high maneuverability and they can be designed to be very robust to the environment they are placed in. The use of quadrupeds as service robots is being realized by the community. Understanding the dynamics of a quadruped robot is a complex task as there are many factors that need to be taken into consideration while designing the robot. This project intends to develop an eight degree of freedom (DOF) sprawling-type quadruped robot. Control of high DOF robots can be highly challenging and computationally heavy. This project intends to work on bringing down the complexity of the robot control by decoupling the dynamics of the robot into four separate dynamic systems, i.e one for each leg, allowing us to achieve a decentralized control of the robot. Gazebo simulation of the robot is achieved with a few gait implementations using a decentralized control strategy.
RushiPDeshmukh/DS595-RL-Projects
RushiPDeshmukh/Face-Recognition-and-Pose-Classifiaction
RushiPDeshmukh/Motion_Planning_Projects
RushiPDeshmukh/PandaPickAndPlace
RushiPDeshmukh/PPO_DQN_CarRacing
Implemented PPO with clipping objective on a continous action space. Modified DQN in discrete action space. To play car racing, an environment of openGym.
RushiPDeshmukh/ROS2Planner
ROS2, Pygame, Discreet Planning Algorithms and Sampling Algorithms.
RushiPDeshmukh/VIO_Benchmarking
Comparing Different VIO methods on standard open source datasets like EuRoC MAV Dataset.
RushiPDeshmukh/M5_4EncoderMotor_RPI_Interface
RushiPDeshmukh/MLX_RL
RushiPDeshmukh/Motion_Primities_based_Planning
Motion planning for a robot with kinematic constraints as well as the dynamic constraints is tackled using motion primates method. This repository will focus on A* and its variants as well as sampling based algorithms for the planning.
RushiPDeshmukh/Motion_Primitives_Based_Planning
RushiPDeshmukh/OakDLite_Interfaces
This package will provide the ROS2 interfaces for OakD cameras
RushiPDeshmukh/ODM
A command line toolkit to generate maps, point clouds, 3D models and DEMs from drone, balloon or kite images. 📷
RushiPDeshmukh/RushiPDeshmukh
Config files for my GitHub profile.
RushiPDeshmukh/RushiPDeshmukh.github.io
Portfolio Website
RushiPDeshmukh/StandardCplusplus
Standard C++ for Arduino (port of uClibc++)