Contact sensors
63 · Solver and runtime
Three 4-box towers. newton.sensors.SensorContact reads each base box's contact force from the solver's contact_f observable, split by counterpart so column 0 is the ground. With sleeping on, settled islands skip their contact rows, so their forces read zero; sleep_skip_constraints=False keeps solving them.
Cost
Time per step, CUDA graph replay, median of 7 interleaved rounds on one NVIDIA RTX PRO 6000 Blackwell Server Edition; ratios are against the first row. Solver allocation is the GPU memory the solver allocates at construction, before warm-up and graph capture. Scene: the three 4-box towers in each world, timed after 4 s of settling.
| Option | 1 world | 8 worlds | solver allocation, 8 worlds | result in this tutorial |
|---|---|---|---|---|
| no sleeping | 1.48 ms | 1.75 ms | 3 MB | sensor reads 39.2 N |
| enable_sleeping=True | 0.28 ms 0.19× | 0.36 ms 0.21× | 3 MB | sensor reads 0 while asleep |
| sleeping, sleep_skip_constraints=False | 1.39 ms 0.94× | 1.94 ms 1.10× | 3 MB | sensor reads 39.3 N |
What to look at
- Without sleeping the ground carries each 4-box tower: 39.2 N against 39.24 N of weight.
- With sleeping the reading drops to 0 at 0.53 s, when the towers fall asleep.
- sleep_skip_constraints=False keeps 39.3 N while asleep.
Key parameter
SensorContact(model, sensing_bodies=..., counterpart_shapes=...)Request its solver_observable_kinds from solver.observables(...), pass them to step, then call sensor.update(state, contacts, observables=...).
Try changing
- Add counterpart_bodies to split the force between the ground and the box above.
- Read total_force_friction on a pushed tower.
Concept
Contact sensing and sleeping
What it is
SensorContact sums the solver's per-contact forces onto chosen bodies or shapes, optionally per counterpart. Sleeping islands normally skip their contact rows to save work, so contact_f, and the sensor, read zero under them; the contact is still there.
In this scene
Sleeping drops the base reading from 39.2 N to 0; keeping the rows restores it. In the timing below the sleeping towers step 5.3x faster, and keeping their rows gives that back.
Considerations
- Cost
- One world: 1.48 ms per step awake, 0.28 ms asleep, 1.39 ms asleep with rows kept; 8 worlds: 1.75, 0.36 and 1.94 ms.
- Failure modes
- Force thresholds (grasp detection, contact rewards) see a false release when an island sleeps.
- Interactions
- Any wake event restores the rows; contact_f also needs the contacts passed to observables().
- Use it when
- Keep sleep_skip_constraints=True when nothing reads forces on resting objects; set it False, or leave sleeping off, when sensors drive decisions.
The script
cd tutorials && python 63_contact_sensors.py prints each variant's final metric; --record out.json also writes the replay. Every tutorial imports replay.py from the same folder.
# SPDX-FileCopyrightText: Copyright (c) 2026 Skild AI and the Steel contributors# SPDX-License-Identifier: Apache-2.0
"""Contact sensors on Steel contacts, and what sleeping does to their readings."""
import newtonimport warp as wpfrom newton.sensors import SensorContactfrom newton.solvers import SolverSteel
from replay import Recorder
DT = 1.0 / 240.0SUBSTEPS = 8
def build(): builder = newton.ModelBuilder() cfg = newton.ModelBuilder.ShapeConfig(mu=0.7) builder.add_ground_plane(cfg=cfg) for k in range(3): for i in range(4): label = f"base{k}" if i == 0 else f"box{k}_{i}" box = builder.add_body( xform=wp.transform((0.3 * k - 0.3, 0.0, 0.05 + 0.1005 * i), wp.quat_identity()), label=label, ) builder.add_shape_box(box, hx=0.05, hy=0.05, hz=0.05, cfg=cfg) return builder.finalize()
def simulate(rec, label, **options): model = build() solver = SolverSteel(model, **options) pipeline = newton.CollisionPipeline(model) contacts = pipeline.contacts() # one sensor per tower base, split by counterpart: column 0 is the ground plane (shape 0) sensor = SensorContact( model, sensing_bodies="base*", counterpart_shapes=[0], request_contact_attributes=False ) kinds = set(sensor.solver_observable_kinds) | ( {"body_awake"} if options.get("enable_sleeping") else set() ) observed = solver.observables(kinds, contacts=contacts) state_0, state_1, control = model.state(), model.state(), model.control() weight = float(model.body_mass.numpy()[0]) * 9.81
run = rec.variant(label, model) for _ in range(rec.frames(4.0)): # ground force on a base = the weight of its 4-box tower reading = sensor.force_matrix.numpy()[:, 0, 2].mean() awake = observed.body_awake if options.get("enable_sleeping") else None run.frame(state_0, awake=awake, base_force_n=reading, tower_weight_n=4 * weight) for _ in range(SUBSTEPS): pipeline.collide(state_0, contacts) solver.step(state_0, state_1, control, contacts, DT, observables=observed) state_0, state_1 = state_1, state_0 sensor.update(state_0, contacts, observables=observed) run.check(solver)
if __name__ == "__main__": rec = Recorder() simulate(rec, "no sleeping") # sleeping islands skip their contact rows, so their contact forces read zero simulate(rec, "enable_sleeping=True", enable_sleeping=True) # keep the rows of sleeping islands: readings stay, the speed-up shrinks simulate( rec, "sleeping, sleep_skip_constraints=False", enable_sleeping=True, sleep_skip_constraints=False, ) rec.save()# SPDX-FileCopyrightText: Copyright (c) 2026 Skild AI and the Steel contributors# SPDX-License-Identifier: Apache-2.0
"""Record Steel tutorial runs as replay JSON for the tutorials page, or print a summary.
Every tutorial imports this module. ``python NN_name.py`` runs the variants and printsthe final metric of each; ``--record out.json`` also writes body poses, shapes and metricseries that the page replays in 3D."""
import argparseimport base64import jsonimport osimport subprocessimport sysimport time
import newtonimport numpy as npimport warp as wp
from source_hash import source_sha256
_RENDERED_TYPES = { newton.GeoType.PLANE, newton.GeoType.SPHERE, newton.GeoType.CAPSULE, newton.GeoType.ELLIPSOID, newton.GeoType.CYLINDER, newton.GeoType.BOX, newton.GeoType.MESH, newton.GeoType.CONE, newton.GeoType.CONVEX_MESH,}
class Recorder: """Collect the variants of one tutorial; ``fps`` is the replay frame rate."""
def __init__(self, fps: int = 30): parser = argparse.ArgumentParser() parser.add_argument( "--record", metavar="OUT.json", help="write a replay file for the tutorials page" ) parser.add_argument("--device", default=None, help="Warp device, for example cuda:0") self.args = parser.parse_args() if self.args.device: wp.set_device(self.args.device) self.fps = fps self.script = os.path.abspath(sys.argv[0]) self.variants: list[Variant] = []
def frames(self, seconds: float) -> int: """Number of replay frames in ``seconds``, including the first.""" return int(round(seconds * self.fps)) + 1
def variant( self, label: str, model: newton.Model, *, ghost: bool = False, slot: int | None = None, scene: str | None = None, ) -> "Variant": """Start one run; ``ghost`` overlays run ``slot``, ``scene`` groups runs for charts.""" variant = Variant(label, model, self.fps, ghost, slot, scene) self.variants.append(variant) return variant
def save(self) -> None: """Print a summary and, with ``--record``, write the replay file.""" for variant in self.variants: last = variant.metrics[-1] if variant.metrics else {} flag = " (rows dropped)" if variant.overflow else "" print(f"{variant.label:<48} {json.dumps(last)} {variant.wall:.2f} s{flag}") if not self.args.record: return source_sha = source_sha256(self.script) doc = { "script": os.path.basename(self.script), "script_sha256": source_sha, "newton_commit": _newton_commit(), "newton_version": newton.__version__, "warp_version": wp.config.version, "gpu": wp.get_device().name, "fps": self.fps, "variants": [v.to_json() for v in self.variants], } with open(self.args.record, "w") as f: json.dump(doc, f, separators=(",", ":")) print(f"wrote {self.args.record}")
class Variant: """One recorded run: a pose per frame for every body, plus named metric values."""
def __init__(self, label, model, fps, ghost, slot, scene): self.label, self.model, self.fps, self.ghost, self.slot = label, model, fps, ghost, slot self.scene = scene self.poses: list[np.ndarray] = [] self.metrics: list[dict] = [] self.awake: list[list[int]] = [] self.particles: list[np.ndarray] = [] self.overflow = False self.start = time.perf_counter() self.wall = 0.0
def frame(self, state: newton.State, awake=None, particles=None, **metrics) -> None: """Record ``state.body_q``, optional awake flags, particle positions and metric values.""" self.poses.append(state.body_q.numpy().astype(np.float32).copy()) if particles is not None: self.particles.append(particles.numpy().astype(np.float32).copy()) self.metrics.append({k: _round(v) for k, v in metrics.items()}) if awake is not None: self.awake.append([int(x) for x in awake.numpy()]) self.wall = time.perf_counter() - self.start
def check(self, solver) -> None: """Note whether the solver dropped rows; call once after the run.""" try: solver.check_constraint_capacity() except RuntimeError: self.overflow = True
def to_json(self) -> dict: poses = ( np.stack(self.poses) if self.poses else np.zeros((0, self.model.body_count, 7), np.float32) ) return { "label": self.label, "fps": self.fps, "ghost": self.ghost, "slot": self.slot, "scene": self.scene, "bodies": int(self.model.body_count), "frames": int(poses.shape[0]), "poses": base64.b64encode(poses.astype(np.float32).tobytes()).decode(), "shapes": _shape_table(self.model), "metric": self.metrics, "awake": self.awake or None, "particles": self._particles_json(), "wall_s": round(self.wall, 3), "overflow": self.overflow, }
def _particles_json(self): if not self.particles: return None frames = np.stack(self.particles) tris = self.model.tri_indices radius = self.model.particle_radius.numpy() return { "count": int(frames.shape[1]), "frames": base64.b64encode(frames.astype(np.float32).tobytes()).decode(), "radius": round(float(radius.mean()), 5) if radius.size else 0.0, "tris": [int(v) for v in tris.numpy().reshape(-1)] if tris is not None and tris.shape[0] else None, }
def _shape_table(model): types = model.shape_type.numpy() scales = model.shape_scale.numpy() xforms = model.shape_transform.numpy() bodies = model.shape_body.numpy() flags = model.shape_flags.numpy() shapes = [] for i, t in enumerate(types): if int(t) not in _RENDERED_TYPES or not flags[i] & int(newton.ShapeFlags.VISIBLE): continue shape = { "b": int(bodies[i]), "t": int(t), "s": [round(float(v), 5) for v in scales[i]], "x": [round(float(v), 6) for v in xforms[i]], } if int(t) in (newton.GeoType.MESH, newton.GeoType.CONVEX_MESH): mesh = model.shape_source[i] shape["v"] = [round(float(v), 5) for v in np.asarray(mesh.vertices).reshape(-1)] shape["i"] = [int(v) for v in np.asarray(mesh.indices).reshape(-1)] shapes.append(shape) return shapes
def _round(value): value = float(value) return round(value, 5) if np.isfinite(value) else None
def _newton_commit(): if os.environ.get("NEWTON_COMMIT"): return os.environ["NEWTON_COMMIT"] try: root = os.path.dirname(os.path.dirname(newton.__file__)) return subprocess.check_output(["git", "-C", root, "rev-parse", "HEAD"], text=True).strip() except (OSError, subprocess.CalledProcessError): return NoneRecorded with SolverSteel on Newton 1.9.0.dev0, Warp 1.19.0, NVIDIA RTX PRO 6000 Blackwell Server Edition. Replays read the state back every frame, so they are not a speed benchmark. Script · page text
Steel docs and tutorials are licensed under Apache-2.0. Third-party notices. "Skild", "Skild AI" and the Skild logo are trademarks of Skild AI and are not licensed under Apache-2.0.