Model Predictive Control for Robotics & Automation
Faster, safer robots with constraint-aware predictive controllers.
Book a free 30-minute callPhD APC/MPC · CERN · IAV · SkySails
Challenges
- Joint limits and collision constraints violated under high-speed trajectories
- Force control for compliant assembly tasks
- Sim-to-real gap for deployed controllers
- Real-time computation budget on embedded hardware
How We Help
NMPC for Motion Planning
Handle obstacle avoidance, joint constraints, and end-effector force limits in a single optimisation.
Embedded Real-Time MPC
Deploy solvers on DSP/FPGA at kHz rates using explicit MPC or code-generated NMPC.
Learning-Augmented Control
Combine model-based MPC with Gaussian Process or neural residual models for sim-to-real transfer.
Relevant Design Patterns
Problems by type of company
Each page collects the problems one type of company runs into, with what they cost and how to check them.
- Aircraft makersAircraft and flying taxis: flight models that miss flight tests7 problems
- Conveyor and intralogistics makersIntralogistics: pallet patterns, sway and conveyors tuned on site3 problems
- Drone makersDrone makers: flight control and tuning rest on few experts10 problems
- Humanoid and exoskeleton makersHumanoids and exoskeletons: balance, whole-body control, reality6 problems
- Logistics and warehouse operatorsLogistics sites: capped grid connection, robots losing position2 problems
- Mobile robot and driverless vehicle makersMobile robots: fleet planning by rules, charging by threshold3 problems
- Navigation and radar sensor makersNavigation and radar sensors: fusing data from several sources2 problems
- Robot integratorRobot integrators: grip force, on-site tuning, too few engineers3 problems
- Robot makerRobot makers: accuracy, vibration, cycle time, control skills16 problems
- Security and defence technologySecurity and defence technology: control experts are scarce7 problems
- Space companiesSpace: guidance, navigation and attitude control with scarce data8 problems
Book a free 30-minute call
Let's discuss how model-predictive control can address your specific challenges. A 30-minute call is enough to assess feasibility.
Book a free 30-minute call →