Route inspection
Define the start, destination, route and obstacles. Record the map, robot trajectory and any human intervention.
Applications / Mining inspection
Mining inspection and robotics research
We are developing the Q1 quadruped platform and OmindOS together for underground inspection, drawing on metal mine field research into access, equipment inspection and review of results.
Tasks and records
Define what to inspect and record
These tasks provide a starting point for research and joint validation. Each requires defined inspection targets, data collection methods and assessment criteria.
Equipment inspection
Assess image and temperature capture at inspection points such as pump rooms, conveyors and pipelines. Select sensors for the specific task.
Inspection records
Agree on timestamps, locations, equipment identifiers and original images. Route findings to people for review and define any connections to existing systems.
Operating conditions
Validate each environmental constraint
Mine type, operating area and site conditions determine the configuration. Low light, dust, moisture and limited communications each require defined tests and assessment methods.
- GNSS-denied areas
- Assess repetitive tunnels, map changes and recovery after localization loss.
- Low light and occlusion
- Review lighting, observation distance and sensor coverage. Retain original images.
- Dust and moisture
- Assess sensor contamination, enclosure protection, cleaning needs and ground traction.
- Slopes and obstacles
- Check gradients, steps, clearance, turning space and accessible routes under the intended payload.
- Communications and power
- Define communications coverage, response to connection loss, manual takeover and the task energy budget.
- Hazardous conditions
- Where combustible gases, explosive dust, high temperatures or corrosion are present, assess the complete robot and site access requirements separately.
Research foundation / Kuanghai Mingmou
Connecting localization, perception and inspection
The Kuanghai Mingmou project studies semantic mining inspection across multiple platforms. Multi-sensor localization provides maps and position, 3D perception represents the environment, and task planning connects mobility with inspection actions.
Multi-sensor localization
Research combines LiDAR, camera and inertial measurements for point-cloud registration, incremental mapping and trajectory estimation in mine tunnels.
3D reconstruction and perception
Research explores neural radiance fields (NeRF) and geometric priors to reconstruct scenes from incomplete observations.
Semantic understanding and planning
Language instructions and scene semantics help break inspection goals into steps that connect route planning, dynamic obstacle avoidance and task execution.
From a map to an inspection record
Associate inspection targets with map locations, organize movement, observation and recording, then assess how those functions work together along the same route.
This section presents earlier Kuanghai Mingmou research. Q1 mobility, payload, runtime and inspection capabilities require separate validation on the relevant prototype under documented test conditions.
Check Q1 development statusField recordings and research context
These photographs document access routes, ground conditions and equipment in an underground metal mine, providing context for mobility, sensing and communications.
The recording below shows a tracked platform underground. Q1 quadruped development demonstrations are presented separately.
Watch the tracked-platform recording
Site conditions
Low light, wet ground and confined passages place different demands on robot movement, sensor placement and communications.
Recorded platform
The footage shows a tracked platform and helps inform site requirements. It is not evidence of Q1 quadruped performance.
Project assessment
Define the route, gradients, obstacles and inspection targets before selecting sensors, communications and field tests.
In development
Q1 quadruped inspection platform
Q1 is at the engineering prototype stage, with development focused on underground mobility, multi-sensor perception and inspection tasks. The concept rendering illustrates that direction; operating suitability requires complete-system and field validation.
Q1 · In developmentJoint validation
Build evidence on one route first
Agree on the task and conditions, then record results at each stage. Both teams confirm the metrics, number of trials and acceptance criteria.
- 01 / Requirements
Define the route and targets
List the route, equipment locations, site conditions and information to collect.
- 02 / Configuration
Confirm the system and interfaces
Record the platform model, sensors, software versions, communications and responsibilities.
- 03 / Testing
Test tasks and exception handling
Record arrival at targets, data collection, human intervention and responses to conditions such as connection loss or low battery.
- 04 / Assessment
Review results and operating scope
Review original records, task completion, open issues and the scope of the next validation stage.
Common questions
Does the underground footage show Q1?
The underground video shows a tracked platform. The point cloud comes from the Kuanghai Mingmou project. Both provide site and research context; see the Q1 product page for its development status and demonstrations.
What can we work on together now?
We can discuss inspection requirements, assess software integration for existing robots and design test tasks together. Q1 is at the engineering prototype stage, so the scope depends on prototype readiness and site conditions.
Can Q1 be used in coal mines or explosive environments?
A mining demonstration alone cannot establish suitability. The model, hazards, certification and site access requirements must be reviewed before considering a test. Metal mine recordings do not demonstrate explosion-protection compliance.