Ground rover
Four-wheel-drive chassis sized for soft soils and low vegetation, with on-board drivetrain and power management.
Lab-tested
Bringing land back to life.
Robotics and artificial intelligence in the service of demining.
A ground rover and its radar designed in Senegal, complemented by a support drone and a predictive model, to detect before a human life is exposed.
Representing an institution or a competent authority? Read the operating framework
Autonomous ground rover Lab-tested
Moves across the area on its own and probes the ground, with no personnel exposed.
Support drone Acquired
Surveys the area from the air and relays the radio link.
Predictive AI In design
Prioritises the surfaces to be probed and produces a risk map.
Maturity: prototype under development
Designed and developed by a multidisciplinary Senegalese engineering team.
- Dakar, Senegal
- Prototype under development
- A team of nine engineers
01 - The system
Five building blocks, one goal: keep people off suspect ground
Each block carries an explicit validation status. Anything not yet tested is flagged as such.
Support drone
Off-the-shelf aircraft, selected for prior aerial survey of the area, photogrammetric mapping and radio relay to the operator station.
Acquired
Ground-penetrating radar
Transmit/receive antennas designed locally, electronics finalised on a printed circuit board and integrated into the rover. First detection trials completed.
Lab-tested
Predictive mapping
Decision-support model combining past detections and terrain data to prioritise the surfaces to be probed.
In design
Mission supervision
ROS 2 autonomous navigation and operator interface: real-time monitoring, immediate manual override, emergency stop.
Lab-tested
Design characteristics
6 h Target mission endurance
1.4 kW Nominal drive power
5.7 km/h Top speed
0.45 m/s Autonomous navigation, accuracy first
100 % Electric drivetrain
02 - Concept of operation
A mission, in cross-section
The drone surveys, the rover probes, the operator decides. Nobody walks on the area until it has been characterised.
Authorisation
The area and framework are set by the competent authority.
Survey
Drone overflight and mapping of the area.
Coverage plan
A path covering the entire surface.
Acquisition
Autonomous progression and continuous ground probing.
Marking
Geo-referencing of anomalies and immediate avoidance.
Handover
Data delivered to the authority and the operator.
Return
The rover returns to its entry point and the mission closes.
03 - Institutional framework
Our place in the national framework
Diambar SAFER is a support tool. It does not conduct demining operations, does not replace any accredited operator, and only intervenes at the request and under the control of the competent authority.
01
Competent authority
Defines the area, authorises the intervention and sets the mission framework.
02
Accredited operator
Runs the mission, supervises the system and validates every operational decision.
03
Diambar SAFER
Acquires, measures, maps. No irreversible action, no autonomous decision.
04
Handover
Geo-referenced data returned to the authority, which retains ownership and use.
↻ The loop closes on the authority: the system produces information, not decisions.
What the system does
- Autonomously covers a delimited area, following a complete coverage plan.
- Acquires ground-penetrating radar signals and records detected anomalies.
- Geo-references every anomaly and produces a usable map.
- Returns all raw and processed data to the requesting authority.
What the system does not do
- It neither neutralises, moves nor destroys any explosive ordnance.
- It does not declare an area cleared - that qualification belongs to the authority.
- It replaces neither the deminer, nor the accredited operator, nor established procedures.
- It undertakes no mission without prior formal authorisation.
04 - Operating doctrine
Four non-negotiable principles
These principles are design constraints, built into the system architecture rather than added on top.
PRINCIPLE 01
The human operator decides
Every mission is continuously supervised. The operator retains manual override and emergency stop at all times. Autonomy covers movement and acquisition, never decision-making.
PRINCIPLE 02
No automated neutralisation
The system detects, locates and maps. It never acts on explosive ordnance. Handling detected objects is the exclusive remit of certified teams.
PRINCIPLE 03
A detection is not a certainty
Results are indications to be confirmed through established procedures. No area is declared safe on the basis of a system acquisition alone.
PRINCIPLE 04
Aligned with the normative framework
The system is designed to fit within the International Mine Action Standards (IMAS / NILAM) and national procedures. No compliance is claimed prior to formal assessment.
Data sovereignty
Detection and mapping data produced by the system constitutes sensitive information infrastructure. It is owned by the requesting national authority, hosted on infrastructure it controls, and is never transferred to any third party without its written consent. No coordinates of suspected areas are published on this site.
05 - Transparency
Where we actually stand
The project has moved from a paper dossier to a rolling prototype. This page is updated at every milestone, including when a milestone is missed.
Done
- Rover assembled, painted and rolling
- Drivetrain and power supply integrated
- Remote control from the operator case
- Articulated arm and emergency stop integrated
- Radar electronics finalised on a printed circuit board
- Ground-penetrating radar integrated into the rover
- First detection trials completed
- Support drone acquired
In progress
- Qualification: electrical safety, electromagnetic compatibility, mechanical robustness
- Autonomous navigation and coverage planning under ROS 2
- Localisation through GNSS / IMU fusion
- Predictive model: architecture and dataset design
- Support drone integration
To be validated
- Demonstration on controlled ground, the decisive remaining step
- Detection performance: detection and false-alarm rates
- Rover endurance and durability in real conditions
- Expected efficiency gain from the predictive model
No detection performance has been measured under operational conditions to date.
06 - Public presence
Where the project has been presented
The prototype has been shown publicly and before juries since 2025. Every appearance is dated and verifiable.
-
Head of State Special Prize for technological innovation
Dakar
Second entry, with a rolling prototype presented to the jury after an initial paper application.
-
National investors mobilisation day
Dakar
Presentation of the project to national investors.
-
Head of State Special Prize for technological innovation
Dakar
First edition. The project was then at design-dossier stage.
07 - The team
One team, a clear chain of responsibility
Engineering and management students in electrical, mechanical and computer engineering. Every building block of the system has a named owner.
-
Cheikh Tidiane Diop
Project lead - ROS supervision and integration
-
Sokhna Diarra Touré
Coordination - management and finance
-
Brady Blaise Bodio a Bodio
Instrumentation and GPR design
-
Seyni Diatta
Power systems and ROS simulation
-
Sidi Aly Dicko
Instrumentation - ROS perception and localisation
-
Mariama Diop
Artificial intelligence and data analysis
-
Mame Yacine Sarr
Control systems and autonomous driving
-
Pape Birame Thiaw
Mechanical design
-
Mouhamed Seck
Mechanical manufacturing
-
Samy Sambou
Mechanical manufacturing
08 - Supporters
Those backing the project with us
The prototype exists thanks to institutional support, academic supervision and the team’s own financial commitment.
-
Funding
Haut Commandement de la Gendarmerie nationale et de la Justice militaire
Financial support for prototype development.
-
Funding and scientific guidance
École Supérieure Polytechnique de Dakar
Financial support, academic supervision and access to workshops and test facilities.
-
Project members’ contribution
Own funding committed by the team since inception.
09 - Get in touch
Contact
Each request is routed to the relevant contact. Institutional enquiries are handled directly by the project management.
Authorities and institutions
Government services, national mine action bodies, defence and security forces: requests for information, demonstration or project oversight.
Write to the project lead→ cheikhtidianediop2@esp.snFunding and cooperation
Donors, cooperation agencies and foundations wishing to review the project file, governance and maturity trajectory.
Request the project file diambarsafer@gmail.comLaboratories and industry
Collaboration on ground-penetrating radar, autonomous navigation, signal processing or the trial campaign.
Propose a collaboration diambarsafer@gmail.com