Perception and sensor fusion
Visual and sensor data processing for environment understanding. A multisensor approach has been explored in the laboratory prototype; specific camera and sensor support depends on the configuration.
Visual context
Cameras provide the information base for environment assessment. Integration defines available video streams, data formats and computing constraints.
Combining data sources
A multisensor approach has been explored in the laboratory prototype. Cameras, inertial inputs and other sources are selected for the partner’s equipment.
Environment representation
Spatial data is converted into a representation for navigation modules. Development focuses on connecting sensor observations with understandable operator decisions.
Obstacle detection
The prototype explores obstacle detection for motion assistance. Field of view, lighting conditions and detection limits are evaluated for each platform.
Data quality assessment
The roadmap includes assessment of sensor availability and consistency. This is intended to help operators understand when extra attention or a mode change is needed.
Diagnostic views
Integration defines processing outputs for engineering review. The partner team can compare input data, module behavior and test results.
Localization and navigation
SLAM, visual-inertial approaches and sensor fusion for position and motion estimation. GNSS-degraded operation is a development focus and requires validation under specific conditions.
Position and motion estimation
Localization combines observations to estimate platform motion. Methods and accuracy criteria depend on sensors, environment and integration objectives.
SLAM and spatial mapping
SLAM and environment representation have been explored in the laboratory. Further work evaluates repeatability on the partner’s configuration and routes.
Visual-inertial approach
Combining visual and inertial information is one direction for motion estimation. Implementation depends on source synchronization and hardware characteristics.
Navigation with degraded GNSS
We are developing localization approaches for unreliable satellite signals. Performance in a specific environment requires a dedicated validation programme.
Routes and navigation constraints
The roadmap covers routes and spatial constraints. Motion parameters, permitted areas and operator intervention conditions are defined for the partner’s scenario.
Repeatable validation
Localization evaluation is organized around agreed routes, conditions and acceptance criteria. This supports configuration comparisons and systematic product development.
Motion assistance
Obstacle detection, control assistance and motion supervision have been explored in the laboratory prototype. Operating limits and system behavior are evaluated during integration.
Assisted control
The laboratory prototype combines assisted control with environment assessment. Software responsibilities and authority limits are defined with the platform manufacturer.
Motion supervision
Motion state assessment helps compare commands with system behavior. Integration defines available feedback signals and how they are presented to operators.
Obstacle-aware assistance
Perception outputs can support the development of assisted navigation functions. Obstacle responses and operating limits are evaluated on the specific platform.
Operator oversight
Operator oversight is part of the integration approach. Mode switching, action confirmation and intervention are defined by the agreed operating scenario.
Behavior under constraints
The roadmap includes defined responses to unavailable data and changing system conditions. Stop conditions, restrictions and handover behavior are specified during platform analysis.
From prototype to pilot
Autonomy is expanded in stages, from bench evaluation to controlled trials. Each additional mode should be supported by results from the preceding validation stage.
Interfaces and integration
Prototype integrations include MAVLink, an operator interface, video streams and web components. A working prototype module does not imply universal compatibility with every UGV.
Platform interfaces
Prototype integrations explore MAVLink and control-system communication. Supported messages, commands and constraints are documented for the agreed configuration.
Operator tools
Work includes operator-interface interaction and prototype scenarios with QGroundControl. The workflow is adapted to the partner team’s roles and tools.
Video and web components
Video streams and web interfaces connect observations with software modules. Format, latency and access requirements are defined before integration is implemented.
Computing adaptation
Modules are aligned with available computing resources. This gives manufacturers a basis for balancing functionality, power consumption and hardware cost.
Modular deployment
The architectural approach composes functions for the partner’s task. A repeatable integration package is a development objective following configuration validation.
Documentation and support
The project scope defines interface descriptions, configuration records and acceptance materials. Support and update coverage are set by the commercial agreement.