Postgraduate Diploma in All-Weather Perception and Domain Adaptation
Online 800-hour Postgraduate Diploma in All-Weather Perception and Domain Adaptation. 8 modules with ISO 21448:2022 (SOTIF), ISO 26262:2018 and PyTorch. Assigned tutor and open enrolment.
Take the controls.
Run a real engineering trade-off from this field. Move the sliders, watch the model react in real time and beat the challenge.
What you will be able to do
Explain the physical effects of weather on cameras, radar and LiDAR.
Use datasets and benchmarks for adverse conditions.
Generate synthetic bad-weather data with physical validity.
Apply unsupervised domain adaptation techniques.
Design weather-robust multimodal architectures.
Estimate uncertainty and detect out-of-distribution conditions.
Validate perception against SOTIF and the operational design domain.
Who it's for
- Perception engineer looking to specialise in All-Weather Perception and Domain Adaptation.
- Computer vision researcher looking to specialise in All-Weather Perception and Domain Adaptation.
- ADAS data engineer looking to specialise in All-Weather Perception and Domain Adaptation.
- SOTIF validation engineer looking to specialise in All-Weather Perception and Domain Adaptation.
- Machine learning specialist looking to specialise in All-Weather Perception and Domain Adaptation.
- Simulation engineer looking to specialise in All-Weather Perception and Domain Adaptation.
- Graduates in engineering or science seeking to enter the automotive sector.
What you'll work with
- ISO 21448:2022 (SOTIF)
- ISO 26262:2018
- ISO 34503:2023
- Reglamento ONU n.º 157
- ISO/PAS 8800:2024
- UL 4600
- ISO 34502:2022
8 modules · 800 h
1.1Physics of rain, fog and snow on light1.2Attenuation and backscatter in LiDAR1.3Radar behaviour in precipitation1.4Glare, backlight and low light1.5Droplets and dirt on optics1.6Wet road surfaces and specular reflections1.7Operational design domain and environmental conditions1.8Documented ADAS failures in bad weather1.9Weather visibility and its measurement1.10Case study: degradation analysis using rain data
2.1ACDC, BDD100K and Foggy Cityscapes2.2Bad-weather radar datasets such as RADIATE2.3DENSE and Seeing Through Fog2.4Annotation in adverse conditions2.5Per-condition metrics and stratification2.6Geographic and seasonal distribution biases2.7Reference synthetic corruptions2.8Data versioning and traceability2.9Weather metadata and its capture2.10Practical: stratifying a dataset by climate
3.1Physical models of fog and rain in images3.2LiDAR simulation with particles3.3Weather rendering in CARLA3.4Image-to-image translation with CycleGAN3.5Diffusion models for augmentation3.6Domain randomisation3.7Validation of synthetic realism3.8The simulation-to-reality gap3.9Simulating snow accumulation on roads3.10Lab: generating synthetic fog
4.1Covariate and label shift4.2Adversarial feature alignment4.3Self-training with pseudo-labels4.4Adaptive batch normalisation4.5Test-time adaptation4.6Source-free adaptation4.7Continual adaptation and catastrophic forgetting4.8Measuring adaptation gain4.9Adaptation across sensors from different manufacturers4.10Lab: adapting a detector to fog
5.1Camera-radar fusion for bad weather5.2Thermal cameras in low visibility5.3Adaptive sensor weighting5.4Robust 3D detection5.5Semantic segmentation in snow5.6Removal of rain echoes in point clouds5.7Image enhancement prior to perception5.8Computational cost in the vehicle5.9Road condition detection5.10Practical: camera-radar fusion on nuScenes
6.1Aleatoric and epistemic uncertainty6.2Ensembles and Monte Carlo dropout6.3Confidence calibration6.4Detection of out-of-distribution conditions6.5Input quality monitoring6.6Controlled function degradation6.7Thresholds and rejection strategies6.8Integration with the safety monitor6.9Estimation of effective visibility6.10Lab: OOD detector for dense fog
7.1ISO 21448 SOTIF and triggering conditions7.2Design of weather scenarios7.3Climate chamber and track testing7.4Operational domain coverage7.5UN Regulation No. 157 and its ODD7.6Safety argument for bad weather7.7Performance metrics per scenario7.8Validation report7.9Testing in controlled fog and rain facilities7.10Case study: fog test plan
8.1Definition of the ODD and target conditions8.2Selection of real and synthetic data8.3Model baseline8.4Domain adaptation strategy8.5Uncertainty and OOD mechanism8.6Climate-stratified evaluation8.7SOTIF safety argument for the system8.8Limitations analysis8.9Written report and presentation8.10Final review with the tutor
How you'll learn
Fully 100 % online training on the SEIUM campus, with downloadable material, recorded classes and a dedicated technical tutor. No fixed timetables: students progress at their own pace within the intake, which makes it possible to combine with work.
Each module combines theory, guided reading of the original standard and applied exercises. The working software is PyTorch, CARLA, MMDetection3D, Open3D, Weights & Biases, in free, academic or evaluation versions. Assignments are based on real procedures: adapting a detector to fog and rain, evaluating camera-radar fusion in bad weather, and designing a SOTIF validation plan per weather condition. Assessment is continuous per module plus the final integrative project.
Build it end to end
Your final project follows the real engineering workflow, step by step, reviewed by your tutor.
Explain the physical effects of weather on cameras, radar and LiDAR.
Use datasets and benchmarks for adverse conditions.
Generate synthetic bad-weather data with physical validity.
Apply unsupervised domain adaptation techniques.
Design weather-robust multimodal architectures.
Where this takes you
Requirements
- Prior technical training or professional experience in the sector. A university degree is not required.
- Internet connection and a computer capable of running the programme's software.
- Sufficient level of Spanish to follow technical regulatory texts.
Ready to start?
Send your application today. Our admissions team replies within 72 hours with your study plan and funding options.
Frequently asked questions
The programme totals 800 hours spread over 6 to 8 months. That is an average of 10 to 15 hours per week, compatible with a full-time job.
Work is carried out with PyTorch, CARLA, MMDetection3D, Open3D, Weights & Biases. Free, academic or evaluation versions are used, and the tutor indicates the free alternative where one exists.
An SEIUM own diploma upon passing the continuous assessment and the integrative project. It is in-house training, not an official qualification from the Ministry.
Yes. Modules passed in a course or diploma in the same area are recognised towards the advanced programme. Check the recognition policy when enrolling.
Related programmes
Postgraduate Diploma in Runtime Monitoring and Safety Kernels
Online 800-hour Postgraduate Diploma in Runtime Monitoring and Safety Kernels. 8 modules covering ISO 26262:2018, UL 4600 and RTAMT. Assigned tutor and open enrolment.
Postgraduate Diploma in SOTIF+Functional Integration with Safety Cases
Online 800-hour Postgraduate Diploma in SOTIF+Functional Integration with Safety Cases. 8 modules with ISO 26262:2018, ISO 21448:2022 and Ansys medini analyze. Assigned tutor and open enrolment.
Diploma in Risk Analysis of ADAS and ALKS Systems
600-hour online postgraduate diploma in the risk analysis of ADAS and ALKS: UNECE R157, ISO 26262, SOTIF, scenarios, simulation and track validation, and R155 cybersecurity, with…
More in Autonomous, ADAS & Software-Defined Vehicle
All Autonomous, ADAS & Software-Defined Vehicle programmes → · How to apply · Fees & financing · Scholarships