Research Assistant (Multi-UAV Target Search)

National University of Singapore; Singapore; On-site; SGD 39,000 - 58,000; Full time

Job summary

National University of Singapore invites applications for a Research Assistant (Multi-UAV Target Search) in the Mechanical Engineering depart­ment. You will develop algorithms for autonomous drone swarms, test them in simulation, and integrate with GCS-based fusion for robust target localization.

The role requires strong Python/C++ coding, ROS experience, and familiarity with Gazebo, PX4 and PyTorch. You will work on cutting-edge autonomous navigation and learning-based planning in GNSS-denied

Qualifications

• Strong coding skills in Python and/or C++, with ROS/ROS 2 experience.
• Experience with robotics simulation platforms such as Gazebo, Unity-based simulators, Isaac Sim/Lab.
• Knowledge of motion planning, mapping, SLAM, frontier-based exploration, or autonomous navigation.
• Experience with multi-agent systems, swarm robotics, task allocation, or distributed coordination preferred.
• Experience with deep learning or reinforcement learning methods, such as PyTorch, multi-agent reinforcement learning, Transformers, or learning-based planning, is a plus.
• Familiarity with LiDAR/camera-based perception, occupancy mapping, semantic mapping, or object localization is preferred.
• Experience with aerial robots, PX4, MAVROS/MAVSDK, or real-world drone deployment is highly preferred.
• Ability to design robust systems under limited communication, onboard computation constraints, and possible agent failures.
• Good literature review, technical writing, and communication skills.
• Experience publishing papers and mentoring undergraduate or master’s students is preferred.

Responsibilities

• Develop and test algorithms in simulation environments, and integrate them with multi-drone planning and communication modules.
• Deploy the system on real drone hardware.
• Investigate frontier-based exploration, multi-agent task allocation, semantic mapping, robust planning under communication loss, and learning-based decision-making for scalable swarm search.
• Design a Ground Control Station (GCS)-based fusion framework to collect partial observations from multiple drones, merge detections, and report final target positions.

Skills

Python

C++

ROS/ROS 2

Multi-agent systems

Deep learning / reinforcement learning

Technical writing

Communication skills

Mentoring students

Education

Bachelor's degree in Engineering/Computer Science

Tools

Gazebo

Unity-based simulators

Isaac Sim/Lab

PX4

MAVROS/MAVSDK

PyTorch

Transformers

Job description

Job Title: Research Assistant (Multi-UAV Target Search)

University-Level Unit: College of Design and Engineering

Faculty/Department-Level Unit: Mechanical Engineering

Employee Category: Research Staff

Location_ONB: Kent Ridge Campus

Posting Start Date: 29/06/2026

Job Description

This project focuses on developing autonomous multi-drone swarm systems for collaborative search and target localization in low-rise urban environments. The target scenarios involve GNSS-denied conditions, low-light areas, cluttered indoor/outdoor spaces, narrow openings, double-storey buildings, limited communication bandwidth, RF communication losses, and potential drone failures during the mission. The candidate will investigate both conventional robotics pipelines and recent AI-based approaches for robust swarm exploration, coverage planning, task allocation, and communication-aware coordination. Each drone will use onboard LiDAR and camera sensors for mapping, navigation, obstacle avoidance, and target localization. The project will also involve designing a Ground Control Station (GCS)-based fusion framework to collect partial observations from multiple drones, merge duplicated target detections, maintain global search progress, and report final target positions. The candidate is expected to develop and test algorithms in simulation environments, integrate them with multi-drone planning and communication modules, and eventually deploy the system on real drone hardware. The work may include frontier-based exploration, multi-agent task allocation, semantic mapping, robust planning under communication loss, and learning-based decision-making for scalable swarm search.

Qualifications

• Strong coding skills in Python and/or C++, with experience in ROS/ROS 2
• Experience with robotics simulation platforms, such as Gazebo, Unity-based simulators, Isaac Sim/Lab, or similar environments
• Knowledge of motion planning, mapping, SLAM, frontier-based exploration, or autonomous navigation
• Experience with multi-agent systems, swarm robotics, task allocation, or distributed coordination is preferred
• Experience with deep learning or reinforcement learning methods, such as PyTorch, multi-agent reinforcement learning, Transformers, or learning-based planning, is a plus
• Familiarity with LiDAR/camera-based perception, occupancy mapping, semantic mapping, or object localization is preferred
• Experience with aerial robots, PX4, MAVROS/MAVSDK, or real-world drone deployment is highly preferred
• Ability to design robust systems under limited communication, onboard computation constraints, and possible agent failures
• Good literature review, technical writing, and communication skills
• Experience publishing papers and mentoring undergraduate or master’s students is preferred