Why Does This Robot Need Two Brains?
A genuine question I had before I understood the answer: the Pi already runs Linux and ROS2. Why bolt on a whole separate microcontroller too?
Embedded & ADAS engineering
How ADAS actually runs, ECU by ECU.
Deep write-ups on the embedded systems inside modern vehicles — the perception ECUs, real-time firmware, and safety-critical buses that turn camera, radar and LiDAR data into steering and braking. Written by engineers who ship it, not market it.
- Projects & Demos1717 notes on Projects & DemosPrototype builds, weekend hacks, and the first time a demo actually drives itself.
- AUTOSAR & Software Architecture44 notes on AUTOSAR & Software ArchitectureBSW layers, RTE generation, and the ARXML config change that breaks a build six weeks later
- Embedded Firmware44 notes on Embedded FirmwareBare-metal drivers, RTOS scheduling, and the bugs that only show up at -40°C.
- +5 moreAll 8 disciplinesEach with its note count and newest write-up.
On screen
Test-track scenarios
ADAS & Sensor Fusion10Opens 10 notes on ADAS & Sensor FusionKalman filters, camera-radar fusion, and the sensor stack behind lane-keep.
Illustrative footage
- technical notes
- 45
- engineering disciplines
- 8
- engineers writing
- 9
Standards and stacks we work to
What we do. From first sketch to a tested system.
We design embedded, robotics and automotive systems from A to Z, build them with your team, review what your engineers are building, or train them to do it themselves.
- Robotics & autonomous systemsMobile robots and autonomous platforms, from the chassis and motor control up to ROS 2, navigation and a digital twin.What you getSystem architecture & component selectionMotor control & PID tuningROS 2 integrationSimulation & digital twin
From our technical notes: Why Does This Robot Need Two Brains?Explore - ADAS & sensor fusionCamera, radar and LiDAR pipelines, the fusion behind them and the driver-assistance features built on top.Latest of 10 notesAutomated Parking: Why Path Planning Is the Easy PartExplore
- Embedded firmware & IoTBare-metal, RTOS and embedded Linux firmware for connected devices and telematics, from bootloader to field updates.Latest of 4 notesThe Silent Killer: Stack Overflow in a System With No Stack ProtectorExplore
- AUTOSAR & vehicle networksClassic and Adaptive AUTOSAR architecture, and the CAN, CAN FD and Ethernet communication between ECUs.Latest of 4 notesAUTOSAR Classic vs Adaptive: Picking the Right One Isn't About Which Is NewerExplore
- Functional safetySafety concepts and the evidence behind them, from hazard analysis to a reviewed safety case.Latest of 2 notesWe Automated 80% of Our HIL Regression Suite. Here's What Broke First.Explore
- Automotive cybersecurityThreat analysis and security design for connected ECUs and their updates.Latest of 2 notesISO 21434 for Engineers Who Just Want to Ship CodeExplore
- AI & perceptionModels that run on the device: datasets, training, optimisation and deployment on embedded targets.Latest of 3 notesSimulating Rain: Why Our LiDAR Model Lied to Us in CARLAExplore
- Not sure where yours fits?Tell us what you’re building and we’ll say how we can help.Start a project
We can work with you in 4 ways
Engineering you can explore. Platforms, perception and tools for the systems we build.
Covers show product concepts or frames from recorded demos.
- 01BlinkThor RobotBlinkThor Robot is a vehicle-oriented robotics product that our team is developing to bring environment recognition, warning functions and automated parking into one physical system. Its development direction focuses on integrated driver-assistance behavior, including Level 2 ADAS research under operator supervision.
- 02BlinkThor ROS 2 Robot
- 03BlinkThor Auto Parking
- 04BlinkThor AI Fire Detection
- 05BlinkThor AutoLabel AI
- 06BlinkThor SimVehicleApp
- 07BlinkThor HMI Simulator
- 08BlinkThor Factory Zone Monitoring
- 09BlinkThor Vision Inspector
BlinkThor RobotView product
ADAS feature explainer
See the assistance stack react.
A concise simulation of how perception, longitudinal control and lane support cooperate around the driver.
Feature index07 stages
Recommended
- ADAS & Sensor FusionTraffic Sign Recognition: From Pixel to Decision in Three Pipeline Stages
- AUTOSAR & Software ArchitectureAUTOSAR Classic vs Adaptive: Picking the Right One Isn't About Which Is Newer
- AUTOSAR & Software ArchitectureAUTOSAR RTE Generation Errors: A Decoder Ring
- Projects & DemosWhat We Actually Look For When Hiring an Embedded Automotive Engineer
Reference architecture
Sense, fuse, actuate.
The ADAS architecture our notes keep coming back to: camera, radar, LiDAR and ultrasonic sensors feeding a domain controller, and the brake, steering and powertrain ECUs that act on its decisions.
The core disciplines we cover. 45 notes across 8 disciplines.
- 17 notes
Projects & Demos
Prototype builds, weekend hacks, and the first time a demo actually drives itself.
- 4 notes
AUTOSAR & Software Architecture
BSW layers, RTE generation, and the ARXML config change that breaks a build six weeks later
LatestAUTOSAR Classic vs Adaptive: Picking the Right One Isn't About Which Is Newer
- 4 notes
Embedded Firmware
Bare-metal drivers, RTOS scheduling, and the bugs that only show up at -40°C.
LatestThe Silent Killer: Stack Overflow in a System With No Stack Protector
- 3 notes
Automotive Networking
CAN, CAN-FD, and LIN diagnostics that keep a bus from silently dying.
LatestUDS Diagnostics Over DoIP: What Actually Changes When You Move Off CAN
- 10 notes
ADAS & Sensor Fusion
Kalman filters, camera-radar fusion, and the sensor stack behind lane-keep.
- 2 notes
Automotive Cybersecurity
ISO/SAE 21434, rollback protection, and the threat models that keep a bootloader honest.
- 3 notes
AI & Perception Engineering
YOLO models, dataset edge cases, and the simulation gap between synthetic rain and the real thing.
LatestSimulating Rain: Why Our LiDAR Model Lied to Us in CARLA
- 2 notes
Functional Safety
ISO 26262 in practice: FMEAs, ASIL decomposition, and safety cases that survive audit.
LatestWe Automated 80% of Our HIL Regression Suite. Here's What Broke First.
Recent write-ups from the team. Written by engineers who ship it.
Giving MentorPi-A1 a Digital Twin: What Changed the Day the Robot Could See Itself
We didn't build the twin to look impressive in a demo. We built it because I was tired of guessing what the robot "felt" from the outside.
An Apology to the Small Components on the RRC Lite Board
Five parts I walked right past on my way to the big square chip in the middle. They deserved better.
Watching PID Actually Work: A Night With EncoderMotor_PID_Demo
"set" said 1.000. "actual" never quite agreed with it. That disagreement is the whole lesson.
An Afternoon Skipping ROS2 Entirely to Talk to the Board in Raw Hex
No launch files, no nodes, no topics. Just a Type-C cable, a serial terminal, and every byte typed by hand.
From the lab
LiDAR captures the surrounding environment in real time, giving our robotic platform accurate spatial awareness for perception and navigation.
The engineers behind BlinkThorLabs. 9 engineers.
Mr. Bill LamDirector, ADAS & Functional SafetyAcademic degree: Master in Software Engineering (AI)
Mr. Binh PhuongPrincipal Embedded Systems Engineer — IoT & TelematicsAcademic degree: Master in Electronics & Telecom
Mr. Hoang NguyenPrincipal Technical Lead — ADAS Simulation & SDVAcademic degree: Master in Computer Engineering
Mr. Dinh BuiPrincipal Robotics Engineer — Perception & Decision MakingAcademic degree: PhD & MS in Electrical & Computer Eng.
Mr. Trong NguyenSenior ADAS & SDV Software EngineerAcademic degree: B.Eng. Electronics & Telecom
Mr. Kien LeSenior ADAS Perception EngineerAcademic degree: B.Eng. Mechatronics
Mr. Nhan BuiSenior ADAS & SDV Software EngineerAcademic degree: B.Eng. Electronics & Telecom
Mr. Hoang HuynhSenior Embedded Software Engineer — AUTOSAR & CybersecurityAcademic degree: Master in Mechatronics Engineering
Mr. Hung NguyenSenior Embedded Software EngineerAcademic degree: B.Eng. Mechatronics