Key Features

• State-of-the-art performance for both convex (linear dynamics) and nonlinear trajectory optimization problems
• Convenient interface for dynamics and problem definition via TrajectoryOptimization.jl and RobotDynamics.jl.
• Supports generic nonlinear state and control constraints at each time step.
• Supports second-order-cone programs (SOCPs).
• Allows initialization of both state and control trajectories.
• Supports integration up to 4th-order Runge-Kutta methods. Higher-order methods are possible but not yet implemented.
• Supports implicit integration schemes such as implicit midpoint.
• Supports optimization on the space of 3D rotations.
• Provides convenient methods for warm-starting MPC problems.
• Provides efficient methods for auto-differentiation of costs, constraints, and dynamics via ForwardDiff.jl and FiniteDiff.jl.