phyz-rigid

Featherstone articulated rigid-body dynamics for phyz.

The classical O(n) algorithms, operating on phyz-model's Model and State.

Function	Computes	Cost
`aba`	forward dynamics `qdd = f(q, v, τ)`	O(n)
`aba_with_external_forces`	forward dynamics with per-body wrenches	O(n)
`rnea`	inverse dynamics `τ = f(q, v, qdd)`	O(n)
`crba`	joint-space mass matrix `M(q)`	O(n²)
`forward_kinematics`	body transforms into `state.body_xform`	O(n)
`kinetic_energy`, `potential_energy`, `total_energy`	conserved quantities	O(n)

Example

use phyz_math::{GRAVITY, Mat3, SpatialInertia, SpatialTransform, Vec3};
use phyz_model::ModelBuilder;
use phyz_rigid::{aba, total_energy};

let model = ModelBuilder::new()
    .gravity(Vec3::new(0.0, -GRAVITY, 0.0))
    .dt(0.002)
    .add_revolute_body(
        "pendulum",
        -1,
        SpatialTransform::identity(),
        SpatialInertia::new(
            1.0,
            Vec3::new(0.0, -0.5, 0.0),
            Mat3::from_diagonal(&Vec3::new(0.083, 0.0, 0.083)),
        ),
    )
    .build();

let mut state = model.default_state();
state.q[0] = 0.3;

// Semi-implicit Euler: this crate provides the dynamics, you write the loop.
for _ in 0..500 {
    let qdd = aba(&model, &state);
    for k in 0..model.nv {
        state.v[k] += model.dt * qdd[k];
    }
    for k in 0..model.nq {
        state.q[k] += model.dt * state.v[k];
    }
}

let e = total_energy(&model, &state);

This crate provides the dynamics, not the time loop — the example above writes its own integrator. If you want a ready-made stepper, phyz::sim has Simulator with semi-implicit Euler and RK4 solvers, and phyz-guardian adds adaptive time-stepping and conservation monitoring on top of these primitives.

Part of phyz

phyz is an open-source differentiable multi-physics simulation workspace in pure Rust. Each crate is independent — adding this one does not pull in the rest.

Licensed under MIT.

phyz-rigid 0.4.0

phyz-rigid

Example

Part of phyz