Hand Kinematics

class dexrobot_kinematics.hand.LeftHandKinematics(urdf_path: Path | None = None, config_path: Path | None = None)[source]

Bases: HandKinematicsBase

Class for left hand kinematics calculations

class dexrobot_kinematics.hand.RightHandKinematics(urdf_path: Path | None = None, config_path: Path | None = None)[source]

Bases: HandKinematicsBase

Class for right hand kinematics calculations

HandKinematicsBase

class dexrobot_kinematics.hand.base.HandKinematicsBase(handedness: str, config_path: Path)[source]

Bases: object

Base class for hand kinematics calculations

forward_kinematics(joint_angles: Dict[str, float], base_pose: Pose | None = None, frame: str = 'hand', end_effector: str = 'fingerpad') → Dict[str, Pose][source]

Compute forward kinematics for all fingers

Parameters:

  • joint_angles – Current joint angles

  • base_pose – Optional base pose of the hand, used to transform output pose when frame is ‘world’

  • frame – Reference frame (‘hand’ or ‘world’)

  • end_effector – Target frame type (‘fingerpad’ or ‘fingertip’)

Returns:

Dictionary of finger poses in specified frame

inverse_kinematics_finger(finger: str, target_pos: Position, base_pose: Pose | None = None, frame: str = 'hand', end_effector: str = 'fingerpad', initial_guess: Dict[str, float] | None = None) → Tuple[Dict[str, float], bool][source]

Solve IK for a single finger

Parameters:

  • finger – Target finger name

  • target_pos – Target position

  • base_pose – Optional base pose of the hand, used to transform target when frame is ‘world’

  • frame – Reference frame (‘hand’ or ‘world’)

  • end_effector – Target frame type (‘fingerpad’ or ‘fingertip’)

  • initial_guess – Initial joint angles

Returns:

Tuple of (joint angles, success flag)

inverse_kinematics_grasp(finger_targets: Dict[str, Position], base_pose: Pose | None = None, frame: str = 'hand', end_effector: str = 'fingerpad', initial_guess: Dict[str, float] | None = None) → Tuple[Dict[str, float], bool][source]

Solve IK for multiple fingers simultaneously.

Parameters:

  • finger_targets – Target positions for each finger

  • base_pose – Optional base pose of the hand, used to transform target when frame is ‘world

  • frame – Reference frame (‘hand’ or ‘world’)

  • end_effector – Target frame type (‘fingerpad’ or ‘fingertip’)

  • initial_guess – Initial joint angles

Returns:

Tuple of (joint angles for all fingers, success flag)

IK Solver

class dexrobot_kinematics.hand.ik_solver.HandIKSolver(robot: RobotWrapper, handedness: str, target_frame: str, config_path: Path, eps: float = 0.0001, max_iter: int = 200, damping: float = 0.01, dt: float = 1)[source]

Bases: object

A Pinocchio-based IK solver for robotic hand that handles joint limits and finger joint synchronization

solve(target_pos: Position, initial_q: ndarray | None = None, orientation_weight: float = 0.0) → Tuple[ndarray, bool][source]

Solve inverse kinematics

Parameters:

  • target_pos – Target end-effector position

  • initial_q – Initial joint configuration

  • orientation_weight – Weight for orientation constraint

Returns:

Tuple of (final joint configuration, success flag)