338 lines
17 KiB
Matlab
338 lines
17 KiB
Matlab
classdef results < matlab.unittest.TestCase
|
|
properties (Constant, Access = private)
|
|
seed = 1;
|
|
domainSize = [150, 150, 100]; % fixed domain size [X, Y, Z]
|
|
end
|
|
|
|
properties (Access = private)
|
|
% System under test
|
|
testClass = miSim;
|
|
|
|
%% Diagnostic Parameters
|
|
% No effect on simulation dynamics
|
|
makeVideo = false; % disable video writing for big performance increase
|
|
makePlots = false; % disable plotting for big performance increase (also disables video)
|
|
plotCommsGeometry = false; % disable plotting communications geometries
|
|
|
|
%% Scenario Reinitialization
|
|
% Number of extra reinitializations per test case (3 n-values x 4 configs = 12).
|
|
% Order: n3/A_1_alpha, n3/A_1_beta, n3/A_2_alpha, n3/B_1_beta,
|
|
% n5/A_1_alpha, ..., n6/B_1_beta
|
|
% Set inspectScenarios = true to pause after init for manual review.
|
|
% At the keyboard prompt, type the number of reinits needed into
|
|
% the variable 'reinitCount', then 'dbcont' to continue.
|
|
inspectScenarios = false;
|
|
reinit = zeros(1, 12);
|
|
|
|
%% Fixed Test Parameters
|
|
useFixedTopology = true; % No lesser neighbor, fixed network instead
|
|
discretizationStep = 0.5;
|
|
protectedRange = 5;
|
|
collisionRadius = 5;
|
|
sensorPerformanceMinimum = 0.005;
|
|
comRange = 20;
|
|
maxIter = 400;
|
|
initialStepSize = 1;
|
|
% Each row: [minX minY minZ maxX maxY maxZ]
|
|
obstacleCorners = [results.domainSize(1)/2, results.domainSize(2)*5/8, 0, results.domainSize(1)*5/8, results.domainSize(2), 35;
|
|
results.domainSize(1)/3, 0, 0, results.domainSize(1)/2, results.domainSize(2)*3/8, 40];
|
|
barrierGain = 1;
|
|
barrierExponent = 1;
|
|
timestep = 0.5;
|
|
dampingCoeff = 2;
|
|
end
|
|
|
|
properties (TestParameter)
|
|
%% Test Iterations
|
|
% Specific parameter combos to run iterations on
|
|
n = struct('n3', 3, 'n5', 5, 'n6', 6); % number of agents
|
|
config = results.makeConfigs();
|
|
end
|
|
|
|
properties (MethodSetupParameter)
|
|
trials = struct('r1', 1, 'r2', 2, 'r3', 3, 'r4', 4, 'r5', 5);
|
|
end
|
|
|
|
methods (TestMethodSetup)
|
|
function setSeed(tc, trials)
|
|
rng(tc.seed + trials);
|
|
end
|
|
end
|
|
|
|
methods (Static, Access = public)
|
|
function c = makeConfigs()
|
|
rng(results.seed);
|
|
abMin = 6; % alpha*beta >= 6 ensures membership(0) = tanh(3) >= 0.995
|
|
alphaDist = rand(1, 2) .* [75, 45];
|
|
betaDist = abMin ./ alphaDist + rand(1, 2) .* [1, 1/8] .* (20 - abMin ./ alphaDist);
|
|
alphaTilt = 10 + rand(1, 2) .* [20, 20];
|
|
betaTilt = abMin ./ alphaTilt + rand(1, 2) .* (50 - abMin ./ alphaTilt);
|
|
sensors = struct('alphaDist', num2cell(alphaDist), 'alphaTilt', num2cell(alphaTilt), 'betaDist', num2cell(betaDist), 'betaTilt', num2cell(betaTilt));
|
|
sensor1 = sigmoidSensor;
|
|
sensor2 = sigmoidSensor;
|
|
sensor1 = sensor1.initialize(sensors(1).alphaDist, sensors(1).betaDist, sensors(1).alphaTilt, sensors(1).betaTilt);
|
|
sensor2 = sensor2.initialize(sensors(2).alphaDist, sensors(2).betaDist, sensors(2).alphaTilt, sensors(2).betaTilt);
|
|
% sensor1.plotParameters;
|
|
% sensor2.plotParameters;
|
|
c = struct('A_1_alpha', struct('objectivePos', [3, 1] / 4 .* results.domainSize(1:2), 'sensor', sensors(1), 'doubleIntegrator', false), ...
|
|
'A_1_beta', struct('objectivePos', [3, 1] / 4 .* results.domainSize(1:2), 'sensor', sensors(1), 'doubleIntegrator', true), ...
|
|
'A_2_alpha', struct('objectivePos', [3, 1] / 4 .* results.domainSize(1:2), 'sensor', sensors(2), 'doubleIntegrator', false), ...
|
|
'B_1_beta', struct('objectivePos', [[3, 1] / 4 .* results.domainSize(1:2); [3, 1] / 4 .* results.domainSize(1:2) + 25 .* [-1, 1] ./ sqrt(2)], 'sensor', sensors(1), 'doubleIntegrator', true));
|
|
end
|
|
end
|
|
|
|
methods (Test)
|
|
function plot1_runs(tc, n, config)
|
|
% if n == 5 && config.doubleIntegrator == true
|
|
% tc.makePlots = true;
|
|
% else
|
|
% tc.makePlots = false;
|
|
% end
|
|
% Compute test case index for reinit lookup
|
|
nKeys = fieldnames(tc.n);
|
|
configKeys = fieldnames(tc.config);
|
|
nIdx = find(cellfun(@(k) tc.n.(k) == n, nKeys));
|
|
configIdx = find(cellfun(@(k) isequal(tc.config.(k), config), configKeys));
|
|
testIdx = (nIdx - 1) * numel(configKeys) + configIdx;
|
|
|
|
% Determine number of reinitializations
|
|
reinitCount = tc.reinit(testIdx);
|
|
|
|
for reroll = 0:reinitCount
|
|
|
|
% Set up fixed-size domain
|
|
minAlt = tc.domainSize(3)/10 + rand * 1/10 * tc.domainSize(3);
|
|
% Place sensing objective(s) at fixed positions from config
|
|
objectiveMu = config.objectivePos;
|
|
numDist = size(objectiveMu, 1);
|
|
objectiveSigma = [];
|
|
for ii = 1:numDist
|
|
sig = [200, 140; 140, 280];
|
|
if ~mod(ii, 2)
|
|
sig = rot90(sig,2);
|
|
end
|
|
sig = reshape(sig, [1, 2, 2]);
|
|
objectiveSigma = cat(1, objectiveSigma, sig);
|
|
end
|
|
tc.testClass.domain = tc.testClass.domain.initialize([zeros(1, 3); tc.domainSize], REGION_TYPE.DOMAIN, "Domain");
|
|
tc.testClass.domain.objective = tc.testClass.domain.objective.initialize(objectiveFunctionWrapper(objectiveMu, objectiveSigma), tc.testClass.domain, tc.discretizationStep, tc.protectedRange, tc.sensorPerformanceMinimum, objectiveMu, objectiveSigma);
|
|
|
|
% Initialize agents
|
|
agents = cell(n, 1);
|
|
[agents{:}] = deal(agent);
|
|
|
|
% Initialize sensor model
|
|
sensorModel = sigmoidSensor;
|
|
sensorModel = sensorModel.initialize(config.sensor.alphaDist, config.sensor.betaDist, config.sensor.alphaTilt, config.sensor.betaTilt);
|
|
|
|
% Initialize fixed obstacles from corner coordinates
|
|
nObs = size(tc.obstacleCorners, 1);
|
|
obstacles = cell(nObs, 1);
|
|
for jj = 1:nObs
|
|
corners = [tc.obstacleCorners(jj, 1:3); tc.obstacleCorners(jj, 4:6)];
|
|
obstacles{jj} = rectangularPrism;
|
|
obstacles{jj} = obstacles{jj}.initialize(corners, REGION_TYPE.OBSTACLE, sprintf("Obstacle %d", jj));
|
|
end
|
|
|
|
% Place agents in small-x, large-y quadrant (opposite objectives)
|
|
% with chain topology: each agent connected only to its neighbors
|
|
midXY = (tc.testClass.domain.minCorner(1:2) + tc.testClass.domain.maxCorner(1:2)) / 2;
|
|
quadrantSize = tc.testClass.domain.maxCorner(1:2) / 2;
|
|
margin = quadrantSize / 6;
|
|
agentBounds = [tc.testClass.domain.minCorner(1) + margin(1), ...
|
|
midXY(2) + margin(2); ...
|
|
midXY(1) - margin(1), ...
|
|
tc.testClass.domain.maxCorner(2) - margin(2)];
|
|
% Find a fixed altitude where sensor performance passes at ALL
|
|
% corners of the placement bounds (worst-case XY)
|
|
corners = [agentBounds(1,1), agentBounds(1,2);
|
|
agentBounds(2,1), agentBounds(1,2);
|
|
agentBounds(1,1), agentBounds(2,2);
|
|
agentBounds(2,1), agentBounds(2,2)];
|
|
agentAlt = tc.testClass.domain.maxCorner(3) - tc.collisionRadius;
|
|
while agentAlt > minAlt + 2 * tc.collisionRadius
|
|
worstPerf = inf;
|
|
for cc = 1:4
|
|
p = sensorModel.sensorPerformance([corners(cc,:), agentAlt], [corners(cc,:), 0]);
|
|
worstPerf = min(worstPerf, p);
|
|
end
|
|
if worstPerf >= tc.sensorPerformanceMinimum * 10
|
|
break;
|
|
end
|
|
agentAlt = agentAlt - 1;
|
|
end
|
|
chainSpacingMin = 0.7 * tc.comRange;
|
|
chainSpacingMax = 0.9 * tc.comRange;
|
|
collisionGeometry = spherical;
|
|
for jj = 1:n
|
|
retry = true;
|
|
while retry
|
|
retry = false;
|
|
|
|
if jj == 1
|
|
% First agent: random XY within bounds, fixed altitude
|
|
agentPos = [agentBounds(1, :) + (agentBounds(2, :) - agentBounds(1, :)) .* rand(1, 2), agentAlt];
|
|
else
|
|
% Place at 0.7-0.9 * comRange in XY from previous agent, same altitude
|
|
dir = randn(1, 2);
|
|
dir = dir / norm(dir);
|
|
r = chainSpacingMin + rand * (chainSpacingMax - chainSpacingMin);
|
|
agentPos = [agents{jj-1}.pos(1:2) + r * dir, agentAlt];
|
|
end
|
|
|
|
% Check within placement bounds (XY only, Z is fixed)
|
|
if any(agentPos(1:2) <= agentBounds(1, :)) || any(agentPos(1:2) >= agentBounds(2, :))
|
|
retry = true;
|
|
continue;
|
|
end
|
|
|
|
% Check sensor performance threshold; lower altitude if it fails
|
|
if sensorModel.sensorPerformance(agentPos, [agentPos(1:2), 0]) < tc.sensorPerformanceMinimum * 10
|
|
agentAlt = max(agentAlt - tc.collisionRadius, minAlt + 1.1 * tc.collisionRadius);
|
|
agentPos(3) = agentAlt;
|
|
% If we've hit the floor and still failing, widen XY search
|
|
if agentAlt <= minAlt + 2 * tc.collisionRadius
|
|
agentBounds = [tc.testClass.domain.minCorner(1) + tc.collisionRadius, ...
|
|
tc.testClass.domain.minCorner(2) + tc.collisionRadius; ...
|
|
tc.testClass.domain.maxCorner(1) - tc.collisionRadius, ...
|
|
tc.testClass.domain.maxCorner(2) - tc.collisionRadius];
|
|
end
|
|
retry = true;
|
|
continue;
|
|
end
|
|
|
|
% Must be within comRange of previous agent (chain link)
|
|
if jj > 1 && norm(agents{jj-1}.pos - agentPos) >= tc.comRange
|
|
retry = true;
|
|
continue;
|
|
end
|
|
|
|
% Must be BEYOND comRange of all non-adjacent agents (sparsity)
|
|
% for kk = 1:(jj - 2)
|
|
% if norm(agents{kk}.pos - agentPos) < tc.comRange
|
|
% retry = true;
|
|
% break;
|
|
% end
|
|
% end
|
|
% if retry, continue; end
|
|
|
|
% No collision with any existing agent
|
|
for kk = 1:(jj - 1)
|
|
if norm(agents{kk}.pos - agentPos) < agents{kk}.collisionGeometry.radius + tc.collisionRadius
|
|
retry = true;
|
|
break;
|
|
end
|
|
end
|
|
if retry, continue; end
|
|
|
|
% No collision with any obstacle
|
|
for kk = 1:nObs
|
|
P = min(max(agentPos, obstacles{kk}.minCorner), obstacles{kk}.maxCorner);
|
|
d = agentPos - P;
|
|
if dot(d, d) <= tc.collisionRadius^2
|
|
retry = true;
|
|
break;
|
|
end
|
|
end
|
|
end
|
|
|
|
% Initialize agent
|
|
collisionGeometry = collisionGeometry.initialize(agentPos, tc.collisionRadius, REGION_TYPE.COLLISION, sprintf("Agent %d Collision Region", jj));
|
|
agents{jj} = agents{jj}.initialize(agentPos, collisionGeometry, sensorModel, tc.comRange, tc.maxIter, tc.initialStepSize, sprintf("Agent %d", jj), tc.plotCommsGeometry);
|
|
end
|
|
|
|
% Randomly shuffle agents to vary index-based topology
|
|
agents = agents(randperm(numel(agents)));
|
|
|
|
end % reroll loop
|
|
|
|
% Inspect scenario if enabled
|
|
if tc.inspectScenarios
|
|
tc.testClass = tc.testClass.initialize(tc.testClass.domain, agents, tc.barrierGain, tc.barrierExponent, minAlt, tc.timestep, tc.maxIter, obstacles, tc.makePlots, tc.makeVideo, config.doubleIntegrator, tc.dampingCoeff, tc.useFixedTopology);
|
|
fprintf("Test %d (n=%d, config=%s): reinit=%d. Inspect plot.\n", testIdx, n, configKeys{configIdx}, reinitCount);
|
|
fprintf("To add reinits, update tc.reinit(%d) and rerun.\n", testIdx);
|
|
keyboard;
|
|
tc.testClass = tc.testClass.teardown();
|
|
return;
|
|
end
|
|
|
|
% Set up simulation
|
|
tc.testClass = tc.testClass.initialize(tc.testClass.domain, agents, tc.barrierGain, tc.barrierExponent, minAlt, tc.timestep, tc.maxIter, obstacles, tc.makePlots, tc.makeVideo, config.doubleIntegrator, tc.dampingCoeff, tc.useFixedTopology);
|
|
|
|
% Save simulation parameters to output file
|
|
tc.testClass.writeInits();
|
|
|
|
% Run
|
|
tc.testClass = tc.testClass.run();
|
|
|
|
% Cleanup
|
|
tc.testClass = tc.testClass.teardown();
|
|
close all;
|
|
end
|
|
% function AIIbeta_plots_3_4(tc)
|
|
% % test-specific parameters
|
|
% tc.makePlots = true;
|
|
% tc.makeVideo = true;
|
|
% maxIters = 400;
|
|
%
|
|
% configs = results.makeConfigs();
|
|
% c = configs.A_2_alpha;
|
|
% c.doubleIntegrator = true; % make a2alpha into a2beta
|
|
%
|
|
% % Set up fixed-size domain
|
|
% minAlt = tc.domainSize(3)/10 + rand * 1/10 * tc.domainSize(3);
|
|
% tc.testClass.domain = tc.testClass.domain.initialize([zeros(1, 3); tc.domainSize], REGION_TYPE.DOMAIN, "Domain");
|
|
%
|
|
% % Set objective
|
|
% objectiveMu = [tc.domainSize(1) * 2 / 3, tc.domainSize(2) * 3 / 4];
|
|
% objectiveSigma = reshape([215, 100; 100, 175], [1, 2, 2]);
|
|
% tc.testClass.domain.objective = tc.testClass.domain.objective.initialize(objectiveFunctionWrapper(objectiveMu, objectiveSigma), tc.testClass.domain, tc.discretizationStep, tc.protectedRange, tc.sensorPerformanceMinimum, objectiveMu, objectiveSigma);
|
|
%
|
|
% % Set agent initial states (fully connected network of 4)
|
|
% centerPos = [tc.domainSize(1) / 4, tc.domainSize(2) / 4];
|
|
% d = tc.collisionRadius + (tc.comRange - tc.collisionRadius) / 4;
|
|
% agentsPos = centerPos + [1, 1; 1, -1; -1, -1; -1, 1] / sqrt(2) * d;
|
|
% agentAlt = minAlt * 1.5;
|
|
% agentsPos = [agentsPos, agentAlt * ones(4, 1) + rand * 5 - 2.5];
|
|
%
|
|
% agents = {agent, agent, agent, agent};
|
|
% cg = spherical;
|
|
% sensorModel = sigmoidSensor;
|
|
% sensorModel = sensorModel.initialize(c.sensor.alphaDist, c.sensor.betaDist, c.sensor.alphaTilt, c.sensor.betaTilt);
|
|
% agents{1} = agents{1}.initialize(agentsPos(1, :), cg.initialize(agentsPos(1, :), tc.collisionRadius, REGION_TYPE.COLLISION, "Agent 1 Collision Geometry"), sensorModel, tc.comRange, maxIters, tc.initialStepSize, "Agent 1", false);
|
|
% agents{2} = agents{2}.initialize(agentsPos(2, :), cg.initialize(agentsPos(2, :), tc.collisionRadius, REGION_TYPE.COLLISION, "Agent 2 Collision Geometry"), sensorModel, tc.comRange, maxIters, tc.initialStepSize, "Agent 2", false);
|
|
% agents{3} = agents{3}.initialize(agentsPos(3, :), cg.initialize(agentsPos(3, :), tc.collisionRadius, REGION_TYPE.COLLISION, "Agent 3 Collision Geometry"), sensorModel, tc.comRange, maxIters, tc.initialStepSize, "Agent 3", false);
|
|
% agents{4} = agents{4}.initialize(agentsPos(4, :), cg.initialize(agentsPos(4, :), tc.collisionRadius, REGION_TYPE.COLLISION, "Agent 4 Collision Geometry"), sensorModel, tc.comRange, maxIters, tc.initialStepSize, "Agent 4", false);
|
|
%
|
|
% obstacles = cell(1, 1);
|
|
% obstacles{1} = rectangularPrism;
|
|
% obstacles{1} = obstacles{1}.initialize([0, tc.domainSize(2)/2, 0; tc.domainSize(1) * 0.4, tc.domainSize(2), 40],REGION_TYPE.OBSTACLE, "Obstacle 1");
|
|
%
|
|
% % Set up simulation
|
|
% tc.testClass = tc.testClass.initialize(tc.testClass.domain, agents, tc.barrierGain, tc.barrierExponent, minAlt, tc.timestep, maxIters, obstacles, tc.makePlots, tc.makeVideo, c.doubleIntegrator, tc.dampingCoeff, tc.useFixedTopology);
|
|
%
|
|
% % Save simulation parameters to output file
|
|
% tc.testClass.writeInits();
|
|
%
|
|
% % Run
|
|
% tc.testClass = tc.testClass.run();
|
|
%
|
|
% % Cleanup
|
|
% tc.testClass = tc.testClass.teardown();
|
|
% end
|
|
end
|
|
|
|
methods
|
|
function c = obstacleCollisionCheck(~, obstacles, obstacle)
|
|
% Check if the obstacle intersects with any other obstacles
|
|
c = false;
|
|
for ii = 1:size(obstacles, 1)
|
|
if geometryIntersects(obstacles{ii}, obstacle)
|
|
c = true;
|
|
return;
|
|
end
|
|
end
|
|
end
|
|
|
|
end
|
|
end |