Warehouse Robotics Simulation (Project 212)

[robertogl]

Contribute to the discussion by asking and/or answering questions, commenting, or sharing your ideas for solutions to project 212

[machiobrian]

Hey, I'm Machio Brian [Kenya], I'll be working on Project 212

[adityabastapure]

In Project 212
How can we create different warehouse models?
I was able to create in 2D but I don't how to create a model in gazebo?
Also can I create dynamic obstacle in 2D environment ?
And how to implement dynamic obstacle avoidance in 2D environment

[robertogl]

Hi @adityabastapure,

We have several examples with a 3D warehouse-like environment created in Gazebo. This example is one of them and you can find the Gazebo scenario within the Virtual Machine that we ship with the examples. You can start with the 'Gazebo Warehouse Robot with Obstacles' world and modify it as you need.

Regarding obstacle avoidance, we ship an implementation of the Vector Field Histograms (VFH) algorithm as part of the Navigation Toolbox. The name of the function is controllerVFH.

I hope this helps!

[adityabastapure]

mobileRobotPRM difference plaannerPRM matlab
please reply as soon as possible

[adityabastapure]

load exampleMaps
map = binaryOccupancyMap(simpleMap);

start = [5,5];
goal = [20,20];

planner = plannerAStarGrid(map);
path = plan(planner,start,goal)

subplot(1,2,1)
show(map)
hold on

getting error while plotting this

[adityabastapure]

[adityabastapure]

ss = stateSpaceSE2;
sv = validatorOccupancyMap(ss);
load exampleMaps
map = occupancyMap(simpleMap);
sv.Map = map;
sv.ValidationDistance = 0.01;

planner = plannerRRT(ss,sv);

planner.MaxConnectionDistance = 2.5;
% planner.BallRadiusConstant = 5;
planner.MaxNumTreeNodes = 100;
start = [5,5,0];
goal = [20,20,0];

[pthObj,solnInfo] = plan(planner,start,goal)

subplot(1,2,1)
show(map)
hold on

plot(solnInfo.TreeData(:,1),solnInfo.TreeData(:,2),'.-'); % tree expansion
plot(pthObj.States(:,1),pthObj.States(:,2),'r-','LineWidth',2) % draw path

path = pthObj.States(:,1:2)

controller = controllerPurePursuit;
controller.Waypoints = path;
controller.DesiredLinearVelocity = 0.6;
controller.MaxAngularVelocity = 2;
controller.LookaheadDistance = 1;

[venkateshbs]

I will get back to you on this one soon.

[ankur-sriv]

load exampleMaps map = binaryOccupancyMap(simpleMap);

start = [5,5]; goal = [20,20];

planner = plannerAStarGrid(map);
path = plan(planner,start,goal)

subplot(1,2,1) show(map) hold on

getting error while plotting this.

Hi @adityabastapure,

plannerAStarGrid works by default for grids, hence start & goal given currently are considered as grid points whose origin is top-left corner. Hence, plot seems wrong.

I understand that you want to search in the world frame of reference with 'X & Y'.
For that you must call plan(planner,start,goal,'world') instead of plan(planner,start,goal) as given below.

load exampleMaps;
map = binaryOccupancyMap(simpleMap);

start = [5,5];
goal = [20,20];

planner = plannerAStarGrid(map);
path = plan(planner,start,goal. 'world')
show(planner);

[adityabastapure]

ss = stateSpaceSE2;
sv = validatorOccupancyMap(ss);
load exampleMaps
map = occupancyMap(simpleMap);
sv.Map = map;
sv.ValidationDistance = 0.01;
planner = plannerRRTStar(ss,sv);

% planner.MaxConnectionDistance = 2.5;
planner.BallRadiusConstant = 10; ***************************** This parameter is not showing any effect on output
% planner.MaxNumTreeNodes = 100;

start = [5,5,0];
goal = [20,20,0];

[pthObj,solnInfo] = plan(planner,start,goal)

subplot(1,2,1)
show(map)
hold on

plot(solnInfo.TreeData(:,1),solnInfo.TreeData(:,2),'.-'); % tree expansion
plot(pthObj.States(:,1),pthObj.States(:,2),'r-','LineWidth',2) % draw path

path = pthObj.States(:,1:2)

% Calculates Path Length
j = size(path);
j = j(1)
Path_Length = 0;
i = 1;
while i < j
W1 = path(i,:);
W2 = path(i+1,:);
distance = norm(W1 - W2)
Path_Length = Path_Length + distance;
i = i + 1;
end
Path_Length

RRT* should have more optimal solution than RRT
depending upon planner.BallRadiusConstant = 10; ***************************** This parameter is not showing any effect on output

But its exactly reverse RRT is having optimal solution than RRT*
Please help me out fast

[ajay102p]

Could you clarify on what you expect on RRTStar output compared to RRT. Is it path length you are referring to?
You could try the plannerBenchmark feature to find which planner provide the least path length . Also for computing path length, consider creating navPath object, which can easily calculate path length.

[adityabastapure]

for PlannerRRT/PlannerRRTStar
IsPathFound: 0

Is their any threshold radius around goal point to consider sampled node is equivalent to goal point

If algorithm compares sampled node exactly to the goal coordinate the solution will be not found easily

[adityabastapure]

I need to run planner until and unless solution is found

[ajay102p]

Hi @adityabastapure,

The default GoalReachedFcn is nav.algs.checkIfGoalIsReached, which has a tolerance of 0.5 meters. That means the planner will be able to find a path even if the goal point is in 0.5m radius. You could easily customize this function by editing the default function and provide the function handle to the GoalReachedFcn property of plannerRRTStar.

[ozayp]

Hi am ozayo from wits university (South Africa), am interested in robotics and am currently learning BSc in mechanical engineering. I am looking for collaboration in project title: Warehouse Robotics Simulation.

[thokozanimnguni]

hi everyone im Thokozani from wits university (South Africa) and studying towards a (BSc Mechanical engineering). i am currently looking for collaborators in the Warehouse Robotics Simulation project

[Prajwal51122]

Hii everyone, I am Prajwal Ahire from Walchand College of Engineering ( India ). I aminterested in working and collaborating on this project

[HLLMIC031]

Hello, I'm Emilia Skye Hills from the University of Cape Town working on this project. I am also wondering if anyone else got stuck waiting to download the VMWare Workstation due to "Account Verification Pending"?