Autonomous Ground Vehicle (AGV) with Ground Penetrating Radar (GPR) Sensor Fusion to Detect Subsurface Anomalies and Objects Using Machine Learning and Digital Signal Processing for Site Inspection

Construction and utility works face significant safety and financial risks from undetected buried services, voids and other subsurface anomalies. Ground Penetrating Radar (GPR) can identify such features non-invasively, but interpreting large volumes of multi-channel GPR data remains dependent on scarce specialist expertise.

This two-year, full-time MPhil project will develop and validate an automated signal-processing and machine-learning pipeline for GPR data collected by an autonomous ground vehicle. The enterprise partner, SSE Analytics Limited, will provide access to an operational autonomous platform equipped with a 16-antenna GPR array and Real-Time Kinematic positioning. The project is centred on data analysis and software development with limited mechanical or electronic redesign of the vehicle.

The student will characterise noise, motion artefacts and variability in GPR B-scan data collected across construction sites; develop preprocessing methods such as denoising, background removal, migration and georeferencing; and design machine-learning models to detect, localise and classify features including buried utilities, voids and compaction anomalies. Where labelled field data are limited, synthetic-data generation or augmentation may be investigated. Performance will be evaluated using detection accuracy, localisation error and false-positive rate, and benchmarked against interpretation by an experienced GPR specialist.

During months 13-15, the student will complete a two- to three-month placement with SSE Analytics, supporting live AGV/GPR surveys, collecting data across varied site and ground conditions, refining the analytics pipeline and comparing its outputs with commercial interpretation.

Expected outputs include a characterised and labelled dataset, a reproducible preprocessing pipeline, validated machine-learning models, an independent benchmarking study, an MPhil thesis, and research suitable for conference or journal publication. The research aims to support faster, safer and more scalable subsurface inspection for construction and utility applications.

Mandatory requirements:

1. Academic qualification
Hold, or expect to hold before registration, a minimum second-class honours degree, Grade 2 or higher (2.2; NFQ Level 8), in a relevant discipline.
Relevant disciplines include electronic, electrical, mechanical, mechatronic, automation, robotics, computer, civil, geotechnical or geomatics engineering; computer science; artificial intelligence; data science; applied physics; geophysics; or another closely related discipline.
Applicants may alternatively be eligible through one of the formal transfer routes permitted under the TU Dublin Graduate Research Regulations. The equivalence of qualifications obtained outside Ireland will be determined by the TU Dublin Graduate Research School.

2. English-language proficiency
Meet the English-language requirements applying to TU Dublin research programmes at the time of admission. IELTS 6.0 overall with no component below 6.0 where evidence of English-language proficiency is required. If required, the applicant must provide results from a recognised test taken within the previous two years. Applicants who have successfully completed an honours bachelor’s degree at Level 8 or a master’s degree at Level 9 through the medium of English may be exempt, subject to verification by the Graduate Research School.

3. Quantitative and technical foundation
Demonstrate, through academic modules, a final-year project, employment or other relevant work, a foundation in one or more of the following:
- Digital signal processing.   
- Machine learning or artificial intelligence
- Data analysis or scientific computing
- Image processing or computer vision
- Robotics, sensing or instrumentation
- Geophysics, GPR or non-destructive testing

4. Programming and data analysis
Demonstrate practical programming or computational data-analysis experience using Python, MATLAB, C/C++ or an equivalent technical environment. The applicant must be capable of developing, testing and documenting software used to process experimental datasets

5. Research and communication capability
Demonstrate analytical problem-solving ability, effective written and verbal communication, and the capacity to work both independently and collaboratively within a multidisciplinary university-enterprise research team.

6. Project participation
Be available to undertake the MPhil on a full-time basis and participate in:
- University-based research and training
- Data-collection and validation activities
- Construction-site visits, subject to applicable health, safety and reasonable-accommodation arrangements
- A two- to three-month placement with SSE Analytics during months 13-15
- Travel associated with the enterprise placement and project fieldwork


Desirable Requirements:

1. GPR, sensing or signal-processing experience
Experience with GPR, radar, non-destructive testing, geophysical sensing or other complex sensor datasets; or experience with signal-processing methods such as filtering, denoising, background removal, migration or feature extraction.

2. Machine learning and computer vision
Experience developing or evaluating machine-learning models for classification, object detection, segmentation, localisation or image-based analysis, including experience with frameworks such as PyTorch, TensorFlow, Keras or scikit-learn.

3. Robotics, positioning or sensor fusion
Experience with mobile robots, autonomous vehicles, ROS or ROS 2, RTK-GNSS/GPS, localisation, mapping, geospatial data or the fusion of measurements from multiple sensors.

4. Research and technical practice
Experience with one or more of the following:
- An undergraduate or postgraduate research project
- Experimental design and statistical evaluation
- Preparation or annotation of technical datasets
- Linux, Git or reproducible software-development workflows
- Research data management
- Technical report writing or academic publication
- Construction-site, surveying or industrial health-and-safety procedures

 Funding Agency : ARISE

Student Stipend per annum

€ 25,000

 Materials & Travel Budget per annum

€ 3,000 (Project costs: €2,000; Travel costs: €1,000; Additional first-year only funds for a laptop: €1,500)

 Fees covered by the funding per annum

€5,500

Duration of Funding

24 months

If you are interested in submitting an application for this project, please complete an Expression of Interest.

 https://forms.office.com/e/0hCcrv2Gkp