# accOpt **Repository Path**: paddleli/accOpt ## Basic Information - **Project Name**: accOpt - **Description**: Online time optimal trajectory generation for bounded vel and acc - **Primary Language**: Matlab - **License**: BSD-3-Clause - **Default Branch**: master - **Homepage**: None - **GVP Project**: No ## Statistics - **Stars**: 0 - **Forks**: 1 - **Created**: 2022-03-27 - **Last Updated**: 2022-03-27 ## Categories & Tags **Categories**: Uncategorized **Tags**: None ## README # Online Trajectory Generation Online trajectory generation with bounded velocity and acceleration. Works for multiple direction case. Dependency: Robotics Toolbox for MATLAB (Peter Corke) http://petercorke.com/wordpress/toolboxes/robotics-toolbox ## Instruction on running demo in MATLAB: ### Rename folder to `accOpt` the folder structure should look like . ├── accOpt │ ├── acc_optimal │ ├── RobClass │ ├── Utils │ ├── judp.m │ ├── LICENSE │ └── README.md ### Add path >> addpath(genpath('./accOpt')); ### Construct robot object for visualization >> rob = RobVizLRMate; ### Initialize visualization >> rob.InitPlot; ### Run trajectory generation simulation determine maximum linear velocity `Vm`, maximum linear acceleration `Am`, and sampling time `dt` >> rob.anim( Vm, Am, dt ); e.g. Vm=1 m/s, Am=10 m/s^2, dt=1/30 s, use `rob.anim(1,10,1/30);` then drag blue ball using mouse. The robot should try to reach the blue ball with Vm and Am in real time. Orientation will be determined automatically.