Gym mecanum kinematics matching to Webots roller-hinge proto

Mecanum proto rewrite in b3cf990 made the wheels truly omnidirectional
in Webots, but with asymmetric slip: forward command produces ~89% of
textbook speed while strafe produces only ~38% plus a consistent
~28% backward bleed-through. v1 BC/RL trained on perfect mecanum
gym kinematics could not herd the new dynamics. To unblock that:

* `mecanum_kinematics_step` gains two parameters that scale the
  realised motion to match a deployed-platform calibration:
    - strafe_efficiency  ∈ (0, 1]  default 1.0
    - strafe_to_forward_bleed     default 0.0
  Forward motion is untouched (textbook X-pattern continues to apply
  to vx_body); only the lateral channel is scaled and bleed is added.
* `RobotConfig` exposes both as drive-config fields with the same
  pass-through defaults so existing diff-drive code and existing
  mecanum training pipelines see no behaviour change.
* `HERDING_MEC_WEBOTS` preset bakes in the values measured against the
  current Webots mecanum proto (strafe_efficiency=0.4,
  strafe_to_forward_bleed=-0.28). Training mecanum BC/RL with this
  preset produces policies that compensate for the imperfect
  physical mecanum at deploy.
* `HerdingEnv` plumbs `RobotConfig.strafe_*` through to
  `mecanum_kinematics_step` so the preset takes effect.
* tools/gen_mecanum_wheels.py is added so the proto's 32 roller
  hinges can be regenerated by editing a single set of constants
  rather than hand-editing 1500+ lines of VRML.

Tests:
* 4 new mecanum_kinematics_step tests (default pass-through, strafe
  scaling, backward bleed, forward unaffected by strafe params).
* 3 new RobotConfig tests (defaults, validation, preset shape).
* Sanity check: gym strafe with HERDING_MEC_WEBOTS over 100 steps
  reproduces the Webots calibration to 2 decimal places.

126 unit tests pass (was 120).

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>

tests/test_diffdrive.py

@@ -127,6 +127,45 @@ def test_mecanum_kinematics_pure_strafe():
     assert math.isclose(y, expected_vy * DT, rel_tol=1e-6)
 
 
+def test_mecanum_kinematics_strafe_efficiency_scales_y():
+    # With strafe_efficiency=0.4, realised strafe should be 40% of ideal.
+    w_fl, w_fr, w_rl, w_rr = -10.0, 10.0, 10.0, -10.0
+    x, y, _ = mecanum_kinematics_step(
+        0.0, 0.0, 0.0, w_fl, w_fr, w_rl, w_rr, WHEEL_R, LX, LY, DT,
+        strafe_efficiency=0.4,
+    )
+    ideal_vy = (-w_fl + w_fr + w_rl - w_rr) * WHEEL_R / 4.0
+    assert math.isclose(y, 0.4 * ideal_vy * DT, rel_tol=1e-6)
+    assert x == pytest.approx(0.0, abs=1e-9)
+
+
+def test_mecanum_kinematics_strafe_bleed_pushes_backward():
+    # Negative bleed means strafe commands also push the body backward.
+    w_fl, w_fr, w_rl, w_rr = -10.0, 10.0, 10.0, -10.0
+    x, y, _ = mecanum_kinematics_step(
+        0.0, 0.0, 0.0, w_fl, w_fr, w_rl, w_rr, WHEEL_R, LX, LY, DT,
+        strafe_efficiency=1.0,
+        strafe_to_forward_bleed=-0.28,
+    )
+    ideal_vy = (-w_fl + w_fr + w_rl - w_rr) * WHEEL_R / 4.0
+    assert math.isclose(y, ideal_vy * DT, rel_tol=1e-6)
+    expected_x = -0.28 * abs(ideal_vy) * DT
+    assert math.isclose(x, expected_x, rel_tol=1e-6)
+
+
+def test_mecanum_kinematics_forward_unaffected_by_strafe_params():
+    # Forward command should be untouched by strafe_efficiency / bleed.
+    w_fl = w_fr = w_rl = w_rr = 10.0
+    x, y, _ = mecanum_kinematics_step(
+        0.0, 0.0, 0.0, w_fl, w_fr, w_rl, w_rr, WHEEL_R, LX, LY, DT,
+        strafe_efficiency=0.4,
+        strafe_to_forward_bleed=-0.28,
+    )
+    expected_vx = (w_fl + w_fr + w_rl + w_rr) * WHEEL_R / 4.0
+    assert math.isclose(x, expected_vx * DT, rel_tol=1e-6)
+    assert y == pytest.approx(0.0, abs=1e-9)
+
+
 def test_mecanum_kinematics_pure_rotation():
     # Pure rotation: vx_body=0, vy_body=0, omega>0.
     # w_fl=-10, w_fr=10, w_rl=-10, w_rr=10 → all sums cancel except omega.