Week 51
17/11/2026
Robust Driving at 15 km/h: Clean Data, the Right Camera Height and a Stable Recipe
Programming Robot Intelligence
17/11/2026
Robust Driving at 15 km/h: Clean Data, the Right Camera Height and a Stable Recipe
10/11/2026
Distillation Confirmed, and New Data for the Corners Where Models Fail
03/11/2026
Seed Independence in Ideas and Four Upstream Contributions to Behavior Metrics
27/10/2026
Behavior Metrics: Right-Lane Percentage and Steering Zero-Crossings per Second — First PilotNet Run in Town02
20/10/2026
Behavior Metrics: Right-Lane Percentage and Steering Zero-Crossings per Second — First PilotNet Run in Town02
13/10/2026
Seed Variance: Deterministic Training, EMA, and Ensembles — First Online Evaluation of 200 Models
06/10/2026
CUDA Training Randomness and Behavior Metrics: First Successful Run with CARLA in Docker
29/09/2026
Driving Metrics Framework, Best Model to Date, and First Steps with Behavior Metrics
22/09/2026
PilotNet PyTorch vs TensorFlow: Root Cause Resolved — The Model Now Drives on Par with the Reference
15/09/2026
PilotNet PyTorch vs TensorFlow: Root-Cause Diagnosis and Lateral Recovery Fix
28/08/2026
Finally lane recovery!
21/08/2026
New approach for DAgger and recovery
07/08/2026
Starting with robustness work
24/07/2026
Testing ramdomness seed impact on datset and models
17/07/2026
Increasing model performance by augmenting dataset
10/07/2026
Getting good results and encapsulating the problems
26/06/2026
Putting what has been learned into practice: increasing samples, croping frames and noise injection
12/06/2026
Fine‑Tuning or a New Model? Finding the Optimal Dataset Distribution
05/06/2026
Dataset Balancing & Generalisation: Quantity vs. Quality, Left-Right Symmetry, and Town02 Surprises
29/05/2026
Getting the best autonomous driving to date
08/05/2026
Exploring New Approaches: Horizontal Crop Augmentation & Comparative Evaluation of Balancing Strategies
30/04/2026
Experimental Design: Six Datasets (Week 30)
24/04/2026
Attacking the “Center Lane Bias”
17/04/2026
Exploring dataset for centering on lane
17/04/2026
Exploring dataset for centering on lane
07/04/2026
Dataset completetion for right lane driving
27/03/2026
Getting the first circuit completation on Town07!
20/03/2026
New dataset ideas
13/03/2026
Training for soft turns
09/03/2026
Dataset construction based on softw turns: Town 07 and Town 12
24/02/2026
Training and first testing under new paradigm for datasets
17/02/2026
Datasets: some sources of problems for training
10/02/2026
Autonomous driver oscillations: dataset analysis
03/02/2026
Dataset Balance Analysis and Latency Issue Investigation
27/01/2026
New Dataset Rebalancing and Training with Refined Data Strategy
20/01/2026
Focusing on lane follower with new indications
13/01/2026
Testing dataset balancing with 3, 5 and 7 bins
09/01/2026
Enhancing autonomous driving robustness through DAgger-based corrective dataset generation
30/12/2025
Dataset Refinement and Comparative Evaluation of MobileNet and PilotNet
23/12/2025
Testing with PilotNet
16/12/2025
Dataset Refinement and Error Correction in Autonomous Steering Control
09/12/2025
Dataset Expansion with DAGGER Philosophy and Imitation Learning Performance Evaluation
02/12/2025
Initial DAGGER Steering Implementation and Autonomous Driving Characterization
25/11/2025
Improved simulation stability and advanced dataset aggregation for robust autonomous driving
18/11/2025
Technical Challenges and Multi-Town Dataset Expansion
11/11/2025
Dataset Expansion and Training Pipeline
04/11/2025
Dataset construction and Initial Model Training
28/10/2025
Establishing the MobileNet Training Pipeline
21/10/2025
Preparing the Training Environment for MobileNet
14/10/2025
New Research Direction: MobileNet Architecture
07/10/2025
New week, new learnings, and more progress...
30/09/2025
Second week on my initial schedule for studying autonomous vehicles!
23/09/2025
Already finishing my first week as a doctoral student. I’m excited and ready for the challenge!