01
Overview
During my May–August 2026 KCL internship, I extended the radiation-monitoring robot into Manual Mode and human-supervised Intelligence Mode. Leading three Summer School students, I developed control and planning workflows, refined the mounting system, and supported repeatable testing and research demonstrations.
02
What was developed
- Manual Mode: browser control with arm, stop, disarm and deadman safeguards.
- Intelligence Mode: Gaussian-process modelling proposes measurements for human approval.
- Integrated live camera, LiDAR, localisation and radiation-survey visualisation.
- Redesigned mounting hardware and produced photo-led setup documentation.
Technical details
The planner considers mapped free space, routes and proximity, with per-waypoint approve, skip and abort controls. Testing included synthetic-source, automated, bench and staged robot-drive checks. I allocated research, software, testing and documentation tasks around students’ strengths, and supported KCL Open Day demonstrations.
Weak-source testing demonstrated the connected workflow. Intelligence Mode remains human-supervised; strong-source performance, clinical accuracy and deployment readiness are not established.
Hardware, control and mapping.
A mobile platform carries the sensing payload.
03
Key technical highlights
- Human-approved measurement planning
- Spatial mapping & uncertainty
- Research supervision & demonstration
04
Tools and skills
Python · Go2 SDK · Gaussian processes · LiDAR · Prototyping