RobotLocomotion/drake
Drake
Drake is a robotics simulation framework and control system modeling tool used for designing, simulating, and verifying the dynamics of complex robotic systems. It functions as a multibody dynamics simulator and a mathematical optimization library, providing a suite of algorithms for trajectory optimization and the simulation of articulated robots.
The framework is distinguished by its block-diagram system for composing dynamical subsystems and its ability to formulate and solve diverse mathematical programs, including linear, quadratic, and nonconvex nonlinear problems. It supports specialized robotic capabilities such as bimanual motion planning and hydroelastic contact simulation to model interactions between flexible bodies and fluids.
The system covers a broad range of engineering capabilities, including model-based design, kinematic tree representation, and rigid body contact simulation. It incorporates numerical integration engines for system dynamics and provides tools for robotic model parsing and visualization.
The core is implemented in C++ with wrappers for Python and Julia to facilitate model design and verification.
Features
- Multibody Dynamics Simulators - Provides a high-fidelity multibody dynamics engine for simulating articulated robotic systems and their physical interactions.
- Robotic Physics and Sensor Simulators - Executes physical simulations of multibody systems using various contact models like point and hydroelastic contact.
- Articulated Body Simulators - Simulates articulated rigid bodies by calculating equations of motion and resolving contact constraints.
- Physics and Simulation - Represents physical systems consisting of interconnected bodies to analyze and verify mechanical interactions.
- Optimization Libraries - Offers a toolkit for formulating and solving linear, quadratic, and nonlinear optimization problems for trajectory planning.
- Robotics Simulators - Models and simulates the physics, dynamics, and contact interactions of articulated robotic bodies in a virtual environment.
- Robot Model Importers - Loads robot descriptions and 3D meshes from standard XML or model files to define physical structure and dynamics.
- Robotic Model Parsers - Converts robotic system descriptions from various input formats into a digital representation for simulation and analysis.
- Model-Based Robotic Design - Creates and verifies the behavior of complex robotic systems through a programmatic, model-based design interface.
- Control Systems - Provides a block-diagram framework for composing dynamical systems and simulating closed-loop robotic controllers.
- Motion Planning Toolkits - Calculates optimal, collision-free trajectories and coordinates complex movements for single or multi-arm robotic systems.
- Constrained Trajectory Optimization - Calculates optimal, collision-free, and kinematically valid paths using constrained dynamical system algorithms.
- Python-C Interfaces - Builds a high-performance C++ core with optimized Python and Julia interfaces for flexible model design.
- Block-Diagram Compositions - Combines multiple dynamical subsystems into a block diagram by connecting input and output ports for system composition.
- Block-Diagram System Composition - Provides a block-diagram system for composing complex dynamical subsystems into a directed graph for simulation.
- Dynamic System Simulators - Executes numerical integration of dynamical systems over time using variable-step solvers and event detection.
- Mathematical Optimization Solving - Formulates and solves linear, quadratic, and nonlinear optimization problems to find optimal control policies.
- Mathematical Program Solvers - Formulates and solves linear, quadratic, and nonlinear optimization problems including custom gradient specifications.
- Model-Based Design - Designs and verifies complex robotic systems by composing dynamical subsystems into block diagrams for analysis.
- Multibody Dynamics Modeling - Defines physical structures and kinematic trees from model files to analyze the motion of interconnected mechanical systems.
- Numerical Integration Engines - Executes system simulations over time using variable-step solvers and event detection to maintain physical accuracy.
- Dynamical System Optimization - Applies optimization algorithms to find optimal trajectories or control policies for complex robotic motions.
- Solver-Agnostic Optimization Interfaces - Translates high-level mathematical programs into specific formats required by linear, quadratic, and nonlinear numerical solvers.
- Trajectory Optimization Suites - Provides a comprehensive suite of algorithms for computing collision-free and kinematically valid paths for robotic manipulators.
- Systems Modeling - Constructs robotic systems by combining integrators, physics models, sensors, and controllers into a unified model.
- Kinematic Tree Representations - Defines robot physical structures as hierarchical trees of frames and joints parsed from XML descriptions.
- Dynamical System Definitions - Creates continuous-time or discrete-time system models using symbolic vector fields or custom state update logic.
- Linear and Quadratic Solvers - Calculates direct solutions for linear systems and convex quadratic programs with linear equality constraints.
- Optimization - Defines decision variables, cost functions, and constraint sets to represent mathematical optimization problems.
- Deformable Mesh Simulation - Models physical interaction and deformation between non-convex meshes in fluid-like or elastic environments.
- Simulation State Management - Maintains a deterministic simulation state by controlling time stepping and state updates for repeatable results.
- Multibody Scene Composition - Assembles multiple robot models into an environment by specifying body poses and welding frames to the world.
- Robot Model Visualizations - Renders robot geometry interactively to toggle between visual, collision, and inertia representations.
- Bimanual Motion Planning - Coordinates two robotic arms to maintain a fixed relative pose between end-effectors while avoiding collisions.
- Deterministic State Simulation - Ensures simulation repeatability through deterministic state updates and synchronized time-stepping.
- Closed-Loop Topology Modeling - Combines kinematic tree dynamics with penalty forces to represent mechanically closed systems like linkages.
- Hydroelastic Contact Modeling - Simulates the interaction between flexible bodies and fluids to predict physical contact and response.
- Mixed-Integer Convex Solvers - Handles optimization problems containing binary variables using specialized solvers or branch-and-bound algorithms.
- Nonlinear Optimization Solvers - Finds solutions for general nonlinear optimization problems using gradient-based or specialized non-convex solvers.
- Numerical Optimization Tools - Translates mathematical objectives and constraints into numerical problems to compute optimal solutions.
- Solver Dispatchers - Dispatches decision variables, objectives, and constraints to the most appropriate linear, quadratic, or nonlinear solver.
- Simulation Event Handlers - Triggers specific system responses based on changes in time, state, or input values during simulation.
- محاكاة الروبوتات - Simulates complex robot dynamics.
- Simulation Environments - Simulation of complex robot dynamics.
- Development Frameworks - Toolbox for robot locomotion and control system design.
- Physics Simulation - Planning, control, and analysis toolbox for nonlinear dynamical systems.
