plannerBiRRT

Create bidirectional RRT planner for geometric planning

Since R2021a

Description

The plannerBiRRT object is a single-query planner that uses the bidirectional rapidly exploring random tree (RRT) algorithm with an optional connect heuristic for increased speed.

The bidirectional RRT planner creates one tree with a root node at the specified start state and another tree with a root node at the specified goal state. To extend each tree, the planner generates a random state and, if valid, takes a step from the nearest node based on the MaxConnectionDistance property. The start and goal trees alternate this extension process until both trees are connected. If the EnableConnectHeuristic property is enabled, the extension process ignores the MaxConnectionDistance property. Invalid states or connections that collide with the environment are not added to the tree.

Creation

Syntax

planner = plannerBiRRT(stateSpace,stateVal)

planner = plannerBiRRT(___,Name=Value)

Description

planner = plannerBiRRT(stateSpace,stateVal) creates a bidirectional RRT planner from a state space object, stateSpace, and a state validator object, stateVal. The state space of stateVal must be the same as stateSpace. The stateSpace and stateVal arguments also set the StateSpace and StateValidator properties, respectively, of the planner.

planner = plannerBiRRT(___,Name=Value) sets properties using one or more name-value arguments in addition to the input arguments in the previous syntax. You can specify the MaxConnectionDistance, MaxIterations, MaxNumTreeNodes, and EnableConnectHeuristic properties as name-value arguments.

Properties

StateSpace — State space for planner

state space object

State space for the planner, specified as a state space object. You can use state space objects such as stateSpaceSE2, stateSpaceDubins, stateSpaceReedsShepp, and stateSpaceSE3. You can also customize a state space object using the nav.StateSpace class.

StateValidator — State validator for planner

state validator object

State validator for the planner, specified as a state validator object. You can use state validator objects such as validatorOccupancyMap, validatorVehicleCostmap, and validatorOccupancyMap3D.

StateSampler — State space sampler for sampling input space

stateSamplerUniform object (default) | stateSamplerGaussian object | stateSamplerMPNET object | nav.StateSampler object

Since R2023b

State space sampler used for finding state samples in the input space, specified as a stateSamplerUniform object, stateSamplerGaussian object, stateSamplerMPNET object, or nav.StateSampler object. By default, the plannerBiRRT uses uniform state sampling.

MaxConnectionDistance — Maximum length between planned configurations

0.1 (default) | positive scalar

Maximum length between planned configurations, specified as a positive scalar.

If the EnableConnectHeuristic property is set to true, the object ignores this distance when connecting the two trees during the connect stage.

Example: MaxConnectionDistance=0.3

Data Types: single | double

MaxIterations — Maximum number of iterations

1e4 (default) | positive integer

Maximum number of iterations, specified as a positive integer.

Example: MaxIterations=2500

Data Types: single | double

MaxNumTreeNodes — Maximum number of nodes in search tree

1e4 (default) | positive integer

Maximum number of nodes in the search tree, specified as a positive integer.

Example: MaxNumTreeNodes=2500

Data Types: single | double

EnableConnectHeuristic — Directly join trees during connect phase

false or 0 (default) | true or 1

Directly join trees during the connect phase of the planner, specified as a logical 0 (false) or 1 (true).

Setting this property to true causes the object to ignore the MaxConnectionDistance property when attempting to connect the two trees together.

Example: EnableConnectHeuristic=true

Data Types: logical

Object Functions

plan	Plan path between two states
copy	Create deep copy of planner object

Examples

Plan Path Between Two States Using Bidirectional RRT

Use the plannerBiRRT object to plan a path between two states in an environment with obstacles. Visualize the planned path with interpolated states.

Create a state space.

ss = stateSpaceSE2;

Create an occupancyMap-based state validator using the created state space.

sv = validatorOccupancyMap(ss);

Create an occupancy map from an example map and set map resolution as 10 cells per meter.

load exampleMaps
map = occupancyMap(ternaryMap,10);

Assign the occupancy map to the state validator object. Specify the sampling distance interval.

sv.Map = map;
sv.ValidationDistance = 0.01;

Update the state space bounds to be the same as the map limits.

ss.StateBounds = [map.XWorldLimits; map.YWorldLimits; [-pi pi]];

Create the path planner and increase the maximum connection distance.

planner = plannerBiRRT(ss,sv);
planner.MaxConnectionDistance = 0.3;

Specify the start and goal states.

start = [20 10 0];
goal = [40 40 0];

Plan a path. Due to the randomness of the RRT algorithm, set the rng seed for repeatability.

rng(100,'twister')
[pthObj,solnInfo] = plan(planner,start,goal);

Display the number of iterations taken for the tree to converge.

fprintf("Number of iterations: %d\n",solnInfo.NumIterations)

Number of iterations: 346

Visualize the results.

show(map)
hold on
% Plot start pose and goal pose
plot(start(1), start(2),plannerLineSpec.start{:});
plot(goal(1), goal(2), plannerLineSpec.goal{:});
% Start tree expansion
plot(solnInfo.StartTreeData(:,1),solnInfo.StartTreeData(:,2), ...
    plannerLineSpec.tree{:})
% Goal tree expansion
plot(solnInfo.GoalTreeData(:,1),solnInfo.GoalTreeData(:,2), ...
    plannerLineSpec.goalTree{:})
% Draw path
plot(pthObj.States(:,1),pthObj.States(:,2),plannerLineSpec.path{:})
legend
hold off

Replan the path with the EnableConnectHeuristic property set to true.

planner.EnableConnectHeuristic = true;
[pthObj,solnInfo] = plan(planner,start,goal);

Display the number of iterations taken for the tree to converge. Observe that the planner requires significantly fewer iterations compared to when the EnableConnectHeuristic property is set to false.

fprintf("Number of iterations: %d\n",solnInfo.NumIterations)

Number of iterations: 192

Visualize the results.

figure
show(map)
hold on
% Start tree expansion
% Plot start pose and goal pose
plot(start(1), start(2),plannerLineSpec.start{:});
plot(goal(1), goal(2), plannerLineSpec.goal{:});
plot(solnInfo.StartTreeData(:,1),solnInfo.StartTreeData(:,2), ...
    plannerLineSpec.tree{:})
% Goal tree expansion
plot(solnInfo.GoalTreeData(:,1),solnInfo.GoalTreeData(:,2), ...
   plannerLineSpec.goalTree{:})
% Draw path
plot(pthObj.States(:,1),pthObj.States(:,2),plannerLineSpec.path{:})
legend
hold off

Plan Path Through 3-D Occupancy Map Using Bidirectional RRT Planner

Load a 3-D occupancy map of a city block into the workspace. Specify the threshold to consider cells as obstacle-free.

mapData = load("dMapCityBlock.mat");
omap = mapData.omap;
omap.FreeThreshold = 0.5;

Inflate the occupancy map to add a buffer zone for safe operation around the obstacles.

inflate(omap,1)

Create an SE(3) state space object with bounds for state variables.

ss = stateSpaceSE3([0 220;0 220;0 100;inf inf;inf inf;inf inf;inf inf]);

Create a 3-D occupancy map state validator using the created state space. Assign the occupancy map to the state validator object. Specify the sampling distance interval.

sv = validatorOccupancyMap3D(ss, ...
     Map = omap, ...
     ValidationDistance = 0.1);

Create a bidirectional RRT path planner with increased maximum connection distance and reduced maximum number of iterations. Set EnableConnectHeuristic property to true.

planner = plannerBiRRT(ss,sv, ...
          MaxConnectionDistance = 50, ...
          MaxIterations = 1000, ...
          EnableConnectHeuristic = true);

Specify start and goal poses.

start = [40 180 25 0.7 0.2 0 0.1];
goal = [150 33 35 0.3 0 0.1 0.6];

Configure the random number generator for repeatable result.

rng(1,"twister");

Plan the path.

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

Visualize the planned path.

show(omap)
axis equal
view([-10 55])
hold on
% Start state
scatter3(start(1,1),start(1,2),start(1,3),"g","filled")
% Start tree expansion
plot3(solnInfo.StartTreeData(:,1),solnInfo.StartTreeData(:,2), ...
      solnInfo.StartTreeData(:,3),".-",Color="g")
% Goal state
scatter3(goal(1,1),goal(1,2),goal(1,3),"y","filled")
% Goal tree expansion
plot3(solnInfo.GoalTreeData(:,1),solnInfo.GoalTreeData(:,2), ...
      solnInfo.GoalTreeData(:,3),".-",Color="y")
% Path
plot3(pthObj.States(:,1),pthObj.States(:,2),pthObj.States(:,3), ...
      "m-",LineWidth=2)

References

[1] Kuffner, J. J., and S. M. LaValle. “RRT-Connect: An Efficient Approach to Single-Query Path Planning.” In Proceedings 2000 ICRA. Millennium Conference. IEEE International Conference on Robotics and Automation. Symposia Proceedings (Cat. No.00CH37065), 2:995–1001. San Francisco, CA, USA: IEEE, 2000. https://doi:10.1109/ROBOT.2000.844730.

Extended Capabilities

C/C++ Code Generation
Generate C and C++ code using MATLAB® Coder™.

Version History

Introduced in R2021a

R2023b: Specify Sampling Approach for Path Planning

You can now specify uniform sampling, Gaussian sampling, MPNet sampling, or a custom sampling approach to generate samples for path planning. Use the name, value argument StateSampler to specify the sampling approach.