Research
Four research modules, one integrated stack
Plan4ARI connects natural communication, planning, sensing, control and human-centred assessment. Each module owns one part of the chain from operator intent to robot motion, and hands its results to the next.
How the modules connect
From intent to motion, with trust built in
M1 · Open IPC
Low-code & natural language
Operator intent becomes a verified task plan.
M2 · Open IPC
Sensing & motion planning
Task goals become safe, collision-aware motions.
M3 · Core IPC
Neural motion control
Planned motions are executed robustly in real time.
M4 · Across both layers Trustworthiness Trust is measured and fed back into design choices.
M1Open IPC
Low-Code Programming and Natural Communication
Enable process experts to express robot tasks through language and structured validation instead of detailed code.
Inputs
Operator intent, process constraints, validated domain knowledge and digital-twin feedback.
Outputs
Formal task representations, Behavior Trees and replanning cues for downstream modules.
Connection Its plans depend on sensing updates and provide structured goals to motion planning and control.
M2Open IPC
Intelligent Sensing and Safe Motion Planning
Maintain scene awareness and generate safe robot motions in environments that can change during execution.
Inputs
Radar updates, 3D camera data, task constraints, robot state and digital-twin models.
Outputs
Registered scenes, tracked objects, collision-aware paths and runtime adaptation requests.
Connection It translates task goals into motion alternatives and sends feasible references toward control.
M3Core IPC
Neural Motion Control
Support robust execution under uncertain dynamics, disturbances and changing robot conditions.
Inputs
Planned references, robot state, performance bounds and learned dynamic information.
Outputs
Adaptive control actions, performance monitoring and reusable learned weights.
Connection It executes planned motions and reports execution quality back to monitoring and planning.
M4Across both layers
Trustworthiness and Human-Centric Optimization
Treat trustworthiness as a measurable design variable for industrial human-robot collaboration.
Inputs
Questionnaire responses, interaction-quality measures, task context and system behaviour data.
Outputs
Assessment dimensions, refined metrics and design feedback for the project knowledge base.
Connection Its measurements inform system choices, interface decisions and future optimisation criteria.
See how the modules fit together
The architecture page shows where each module runs, in the Open IPC or the Core IPC, and how commands and feedback move between them.