/static-ipas

Interactive Planning and Sensing for static environments

Primary LanguageMATLABOtherNOASSERTION

The files contained here can reproduce the figures present in the article "Interactive Planning and Sensing in Unknown Static Environments with Task-Driven Sensor Placement"

The script 'main_sim_ipas_v_info.m' runs a single case of an environment with Np parameters using Ns available sensors. It plots the true perfect knowledge field with corresponding optimal path. Then sequentially plots each iteration of the algorithm until the algorithm termination. This is stored as a gif and set of images. Finally a summary plot of the estimated path cost variance, path cost, and nubmer of sensors used in each iteration is plotted. Another similar plot compares the IPAS results with an information based approach.

The script 'gen_sim_data_ipas_v_info_parallel.m' runs a parameter sweep over different grid sizes, number of parameters, and number of available sensors and stores the data in a .mat file. This version uses a 'parfor' loop and requires the MATLAB Parallel Computing Toolbox, however it is easy to modify for standard for-loop computation by changing 'parfor' to 'for' and commenting out lines related to 'parfor_progress'. The script 'ipas_v_info_sim_summary_parellel.m' uses the .mat data from 'gen_sim_data_ipas_v_info_parallel.m' to plot various characteristics of the parameter sweep such as the number of iterations required for an Np-Ns case, comparing variances between IPAS and frame potential information approach, path cost suboptimality, and more.