Degrees of Freedom (DoF)

A Degree of Freedom (DoF) is an independent parameter required to uniquely define the configuration of a mechanical system in space. For a single unconstrained rigid body in 3D space, there are 6 DoF: three for translational position (x, y, z) and three for rotational orientation (roll, pitch, yaw). Mathematically, the system's configuration space is a manifold whose dimension equals the total number of DoF. The physics engine's core task is to solve the equations of motion for these variables over time.

Joints and constraints actively reduce the number of independent DoF in a system. For example:

• A hinge joint removes 5 DoF, leaving only 1 rotational DoF.
• A prismatic (sliding) joint also removes 5 DoF, leaving 1 translational DoF.
• A spherical (ball) joint removes 3 translational DoF, leaving 3 rotational DoF. The constraint solver within the physics engine calculates forces to enforce these restrictions, ensuring the simulated bodies move only along their permitted DoF. The total system DoF equals the unconstrained DoF minus the number of independent constraint equations.

In articulated systems like robotic arms, the total DoF is the sum of the DoF provided by each joint. This is central to forward and inverse kinematics. A common 6-axis industrial robot has 6 DoF, allowing it to position its end-effector in 3D space with a specific orientation. The Articulated Body Algorithm (ABA) and Featherstone's algorithms are designed to compute the dynamics of these multi-DoF chains in O(n) time, making simulation of complex robots efficient. The kinematic tree structure defines how DoF are coupled through parent-child body relationships.

The state vector of a simulated system contains all its DoF and their first-time derivatives (velocities). For a rigid body, this is a 12-dimensional vector: [position (3), orientation (3), linear velocity (3), angular velocity (3)]. The time integration solver (e.g., Symplectic Euler, Runge-Kutta) advances this state vector by solving Newton-Euler equations. The computational cost of a simulation step scales with the number of DoF. Reduced-coordinate models exploit constraints to simulate with a minimal state vector, while maximal-coordinate models (used by many game engines) use a full 6-DoF per body and solve constraints via forces.

For soft bodies and deformable materials, DoF are distributed across the entire object. Simulation methods define these differently:

• Finite Element Method (FEM): DoF are the displacement vectors at each node of a mesh. A complex mesh can have thousands to millions of DoF.
• Mass-Spring Systems: DoF are the positions of each point mass.
• Position-Based Dynamics (PBD): Directly manipulates particle positions (DoF) to satisfy constraints.
• Material Point Method (MPM): Uses Lagrangian particles (DoF) embedded in an Eulerian grid. Simulating these high-DoF systems requires specialized solvers and significant computational power.

In robotics simulation, a subset of a system's DoF are typically actuated (powered by motors or actuators), while others may be passive (like a free-floating base or compliant element). A fully-actuated system has as many independent actuators as DoF, allowing arbitrary force/torque in each. An underactuated system has fewer actuators than DoF (e.g., a quadrotor with 4 actuators for 6 DoF), making control more challenging. Reinforcement learning policies output actions that map to forces or torques applied to these actuated DoF to achieve desired motion.

A single, unconstrained rigid body in three-dimensional space has six degrees of freedom. This is the most fundamental example in physics simulation.

• Three Translational DoF: Movement along the X, Y, and Z axes (position).
• Three Rotational DoF: Rotation about the X, Y, and Z axes (orientation, often represented as roll, pitch, and yaw).

In a simulation engine, this body's state is fully defined by its 3D position (a vector) and its 3D orientation (a quaternion or rotation matrix).

A prismatic joint allows linear translation along a single, fixed axis, connecting two rigid bodies. It provides one degree of freedom.

Real-world examples include:

• A pneumatic piston in a factory.
• A drawer sliding on its rails.
• The linear actuator on a 3D printer gantry.

In simulation, this joint is modeled by a constraint that removes five DoF (all rotations and two translations), leaving only motion along the designated slide axis. The joint's state is defined by a single scalar position value.

A revolute joint allows rotation around a single, fixed axis, connecting two rigid bodies. It is the most common joint in robotics and provides one degree of freedom.

Real-world examples include:

• An elbow or knee joint.
• A door hinge.
• A servo motor in a robotic arm.

Simulation constraints remove five DoF, permitting only angular movement around the hinge axis. The joint's state is defined by a single angular position (an angle).

A typical serial-link industrial robot (e.g., a UR5 or KUKA arm) has six degrees of freedom, matching the DoF of a free body in space. This allows it to position its end-effector (e.g., a gripper) at any arbitrary position and orientation within its workspace.

Configuration:

• Usually comprises six revolute joints in series.
• Total DoF = 6 (1 DoF per joint).

This design provides kinematic dexterity. The system's full configuration is defined by six joint angles (θ₁ to θ₆). In simulation, this is a classic articulated body system solved by algorithms like Featherstone's.

A typical differential-drive wheeled robot operating on a flat plane (like a warehouse floor) has three degrees of freedom in its planning and control model.

These DoF represent its pose on the 2D plane:

• Two Translational DoF: Position (x, y) on the floor.
• One Rotational DoF: Heading or orientation (θ, yaw).

Important Note: Its actuation is often via two independently driven wheels, which creates only two controllable degrees of freedom (forward/backward speed and turning rate). This difference between state DoF and control DoF is a key concept in non-holonomic systems.

A bipedal robot's torso, when considering only the upper body (excluding legs for walking), often has significant DoF for manipulation and balance.

A simplified model might include:

• A 3-DOF waist (yaw, pitch, roll).
• Two 7-DOF arms (similar to a human arm: 3 shoulder, 1 elbow, 3 wrist).
• A 2-DOF neck (pan and tilt).

Total DoF for this subsystem: 3 (waist) + 14 (arms) + 2 (neck) = 19 Degrees of Freedom.

Simulating this requires a sophisticated multibody dynamics solver to handle the complex tree of constraints and compute realistic motion.

The Articulated Body Algorithm (ABA) is an efficient O(n) algorithm for computing the forward dynamics of tree-structured robotic systems or kinematic chains. It recursively propagates inertia and forces through the system's links to calculate accelerations, given applied forces and torques. This algorithm is central to simulating multi-DoF systems like robotic arms in real-time.

• Key Use: Calculating how a robotic arm will move when motors apply torque.
• Efficiency: Linear computational complexity relative to the number of bodies/DoF.
• Relation to DoF: Directly solves for the acceleration of each independent DoF in an articulated system.

Featherstone's algorithm refers to a family of O(n) algorithms for solving the dynamics of articulated multi-body systems. It includes the Composite Rigid Body Algorithm (CRBA) for inverse dynamics and the Articulated Body Algorithm (ABA) for forward dynamics. These methods exploit the sparse, tree-like structure of robotic systems to avoid the O(n³) cost of naive matrix operations.

• Key Use: Efficient simulation of complex robots and biomechanical models.
• Core Methods: CRBA (inverse dynamics), ABA (forward dynamics).
• Relation to DoF: Provides the computational machinery to handle systems with high DoF counts practically.

Multibody dynamics is the study and simulation of mechanical systems consisting of multiple rigid or flexible bodies interconnected by joints and constraints. It provides the theoretical framework for describing systems like vehicle suspensions, robotic manipulators, and humanoid robots. The field's equations of motion directly incorporate the system's degrees of freedom.

• Key Use: Modeling and analyzing complex mechanical assemblies.
• System Types: Can include closed-loop chains, sliding joints, and flexible elements.
• Relation to DoF: The primary goal is to formulate and solve equations governing the evolution of the system's DoF over time.

A constraint solver is an algorithmic component of a physics engine that resolves forces and impulses to satisfy physical constraints between simulated bodies. Constraints mathematically reduce a system's effective degrees of freedom (e.g., a hinge joint removes five DoF). Solvers compute the forces needed to maintain these restrictions, such as joint limits or non-penetration contact.

• Key Use: Enforcing joint connections, contact, and joint limits in simulation.
• Common Methods: Sequential Impulse, Projected Gauss-Seidel (PGS), solving Linear Complementarity Problems (LCP).
• Relation to DoF: Actively restricts the permissible motion of bodies, reducing their independent DoF.

Inverse dynamics is the computational process of calculating the forces and torques required at a system's joints to produce a desired motion or trajectory. Given a kinematic chain's positions, velocities, and accelerations, it solves for the necessary actuator inputs. This is crucial for robot control and motion planning in systems with many degrees of freedom.

• Key Use: Computing motor torques for a robot to follow a pre-planned path.
• Algorithm: Often solved efficiently using the Composite Rigid Body Algorithm (CRBA).
• Relation to DoF: Solves for the control inputs (forces/torques) associated with each actuated DoF.

Forward dynamics is the complementary process to inverse dynamics. It calculates the resulting motion (accelerations) of a physical system when given the applied forces and torques. This is the core computation performed during each time step of a physics simulation to propagate the state of a multi-degree-of-freedom system forward in time.

• Key Use: Simulating how a mechanism will move when forces are applied.
• Algorithm: Solved efficiently by the Articulated Body Algorithm (ABA).
• Relation to DoF: Computes the acceleration for each independent DoF, which is then integrated to update position and velocity.