wzyn20051216/matlab-agent-skills / skills/matlab-robotics-autonomy
matlab-robotics-autonomy
VerifiedMATLAB robotics and autonomy workflow for rigid-body transforms, quaternions, forward/inverse kinematics, trajectory planning, rigid-body dynamics, ROS/ROS 2 bags, sensor fusion, state estimation, SLAM, and reproducible robot experiments. Use whenever the user asks MATLAB to verify a rotation or quaternion, compute FK/IK/Jacobians, check a trajectory against waypoints and limits, check dynamics (M/C/g, energy), or judge whether a recorded state estimate is consistent with its covariance. Missing toolbox, bag, robot model or hardware means SKIP, never a claimed result.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 10.1 KB |
| references/robotics-verification-contract.md | markdown | 17.6 KB |
| scripts/mas_dyn_check.m | other | 11.9 KB |
| scripts/mas_ekf_check.m | other | 6.9 KB |
| scripts/mas_estimate_check.m | other | 13.2 KB |
| scripts/mas_frame_check.m | other | 8.6 KB |
| scripts/mas_kin_check.m | other | 4.5 KB |
| scripts/mas_kin_ik.m | other | 6.1 KB |
| scripts/mas_quat_check.m | other | 12.2 KB |
| scripts/mas_traj_check.m | other | 11.0 KB |
| tests/test_robotics.m | other | 13.5 KB |
| tests/test_robotics_contract.m | other | 25.1 KB |
scripts/mas_kin_check.m
View on GitHubfunction r = mas_kin_check(L, q, opts)
%MAS_KIN_CHECK Planar n-R forward kinematics with the Jacobian cross-checked.
%
% r = mas_kin_check([0.5 0.4 0.3], [0.2; -0.5; 0.1])
%
% r.pos [x; y] tip position
% r.theta absolute angle of each link (cumsum of q)
% r.J analytic Jacobian, 2 x n
% r.Jnum central-difference Jacobian -- different method, same quantity
% r.jacErr max|J - Jnum| / max|J|
% r.singular rank loss (tip at a workspace boundary or at the base)
% r.condJ condition number, capped
% r.manip manipulability sqrt(det(J*J'))
% r.maxReach sum(L), the exact reach of a chain with free-rotating joints
% r.onBoundary |pos| at (or beyond) maxReach
% r.verdict
%
% WHY THE JACOBIAN IS COMPARED AGAINST A FINITE DIFFERENCE
% Every downstream consumer -- resolved-rate control, IK, singularity maps, the
% manipulability ellipsoid, force/torque mapping -- is built on J. A sign error or a
% transposed index in the analytic Jacobian does not throw, does not look wrong on a
% plot of the arm, and produces a controller that drives the tip the wrong way.
% Differentiating the FK numerically shares no code with the analytic derivative, so
% agreement between the two is real evidence and disagreement is a defect. It is
% checked on every call rather than left to the caller.
%
% The analytic form used here: for a planar chain with absolute link angles
% theta_i = sum_{k<=i} q_k,
% dx/dq_j = -sum_{i>=j} L_i sin(theta_i), dy/dq_j = sum_{i>=j} L_i cos(theta_i)
% -- the OUTER sum appears because joint j moves every link from j outward. Dropping
% it is the standard mistake: it yields a Jacobian that is correct for the first joint
% and silently wrong for all the others.
if nargin < 3, opts = struct(); end
opts = dflt(opts, struct('h', 1e-6, 'singTol', 1e-8));
% L, q and theta are COLUMNS throughout. Mixing orientations here silently
% produces an outer product: L(j:n) as a row times theta(j:n) as a column gives a
% K-by-K matrix, sum() then returns a 1-by-K vector, and the assignment into
% J(1,j) fails -- but only for n > 2, so a 2-link smoke test would pass.
L = L(:); n = numel(L);
q = q(:);
if numel(q) ~= n
error('mas_kin_check:size', 'L has %d links but q has %d joints', n, numel(q));
end
if any(L <= 0) || ~all(isfinite(L))
error('mas_kin_check:L', 'link lengths must be finite and positive');
end
[pos, theta] = fk(L, q);
% ---- analytic Jacobian ----------------------------------------------------
J = zeros(2, n);
for j = 1:n
J(1, j) = -sum(L(j:n) .* sin(theta(j:n)));
J(2, j) = sum(L(j:n) .* cos(theta(j:n)));
end
% ---- numerical Jacobian (central difference) ------------------------------
h = opts.h;
Jnum = zeros(2, n);
for j = 1:n
qp = q; qp(j) = qp(j) + h;
qm = q; qm(j) = qm(j) - h;
Jnum(:, j) = (fk(L, qp) - fk(L, qm)) / (2*h);
end
scale = max(abs(J(:)));
if scale == 0, scale = 1; end
jacErr = max(abs(J(:) - Jnum(:))) / scale;
% ---- conditioning ---------------------------------------------------------
s = svd(J);
sMax = max(s); sMin = min(s);
condJ = min(sMax / max(sMin, eps(sMax)), 1/realmin);
singular = sMin < opts.singTol * max(sMax, 1);
manip = sqrt(max(prod(s.^2), 0));
maxReach = sum(L);
onBoundary = norm(pos) > maxReach - 1e-9;
warn = {};
if jacErr > 1e-6
warn{end+1} = sprintf(['analytic and numerical Jacobians disagree by %.2e: one ' ...
'of them is wrong, and every IK / resolved-rate result built on it is wrong too.'], jacErr);
end
if singular
warn{end+1} = ['Jacobian lost rank: the tip is at a workspace boundary or the ' ...
'base, so one Cartesian direction is unreachable.'];
end
r = struct();
r.pos = pos; r.theta = theta; r.L = L; r.q = q;
r.J = J; r.Jnum = Jnum; r.jacErr = jacErr;
r.singular = singular; r.condJ = condJ; r.manip = manip;
r.maxReach = maxReach; r.onBoundary = onBoundary;
r.warnings = warn;
r.ok = isempty(warn);
r.verdict = tern(r.ok, 'PASS', 'FAIL');
r.detail = sprintf('tip (%.4f, %.4f), jacErr %.2e, cond %.3g, manip %.4f -- %s', ...
pos(1), pos(2), jacErr, condJ, manip, r.verdict);
end
% ---------------------------------------------------------------------------
function [pos, theta] = fk(L, q)
theta = cumsum(q(:));
pos = [sum(L(:) .* cos(theta)); sum(L(:) .* sin(theta))];
end
% ---------------------------------------------------------------------------
function o = dflt(o, d)
f = fieldnames(d);
for k = 1:numel(f)
if ~isfield(o, f{k}), o.(f{k}) = d.(f{k}); end
end
end
function s = tern(c, a, b)
if c, s = a; else, s = b; end
end
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| planning-and-task-breakdownaddyosmani/agent-skillsBreaks work into ordered tasks. Use when you have a spec or clear requirements and need to break work into implementable tasks. Use when a task feels too large to start, when you need to estimate scope, or when parallel work is possible. | Stars: 119k | Updated | Claude CodeCodex | Verified | Install |
| screenpipethedotmack/claude-memSearch screen and audio history recorded by Screenpipe and optionally save selected context into claude-mem. Use when the user asks to recall desktop activity, a meeting, or something seen in another app through Screenpipe. | Stars: 118k | Updated | Claude CodeCodex | Verified | Install |
| contact-widgetnexu-io/open-designSelf-contained floating chat widget with welcome screen, social links, meeting button, and message input. Single HTML file, zero dependencies. | Stars: 108k | Updated | Claude CodeCodex | Verified | Install |
| meeting-notesnexu-io/open-designMeeting notes page — title bar with attendees, agenda checklist, decisions block, action items table with owners + dates, and a "next meeting" footer. Use when the brief mentions "meeting notes", "minutes", "1:1 notes", "all-hands recap", or "会议纪要". | Stars: 108k | Updated | Claude CodeCodex | Verified | Install |