Autonomous vehicle - design questions to ponder
Designing an autonomous or remotely controlled vehicle forces tradeoffs between purpose, environment, and terrain. This post lists practical questions to drive decisions about mission profile, ruggedization, power and data handling, sensing, and modularity so your robot survives and completes tasks. Engineers will get a concise checklist to provoke design choices and avoid common failure modes in real-world deployments.
Global Variables vs. Safe Software
10,000 global variables is a striking code smell, and Stephen Friederichs uses the Toyota unintended-acceleration case to show why. He argues the issue was not mere coding style but absent processes: poor testing, lack of enforced standards, incorrect stack analysis, and unowned autogenerated code. Read this for a practical take on why globals and weak processes in safety-critical systems are a much bigger danger than style debates suggest.
The three laws of safe embedded systems
This short article is part of an ongoing series in which I aim to explore some techniques that may be useful for developers and organisations that are beginning their first safety-related embedded project.
Developing software for a safety-related embedded system for the first time
Developing a safety-related embedded product is not the same as writing ordinary firmware, and this article lays out eight practical steps to get you started. Using a washing-machine controller as a running example, it covers scoping, key requirements, hazard analysis, applicable standards, platform and MCU choices, runtime monitoring, and prototyping. The checklist helps teams prepare for verification, testing, and later certification work.
Number Theory for Codes
If CRCs have felt like black magic, this post peels back the curtain with basic number theory and polynomial arithmetic over GF(2). It shows how fixed-width processor arithmetic becomes arithmetic in a finite field, how bit sequences are treated as polynomials, and why primitive polynomials generate every nonzero element. You also get practical insights on CRC implementation with byte tables and LFSRs.
“Smarter” cars, unintended acceleration – and unintended consequences
Smarter cars are arriving fast, but the software tricks behind them may be creating new safety and compliance risks. This post connects Tesla’s autopilot, the VW emissions scandal, and a reported Porsche throttle-delay case to ask whether automotive standards and regulations are keeping pace with increasingly intelligent vehicle control systems.
Coding Step 3 - High-Level Requirements
Stephen Friederichs turns the series toward embedded code by showing how to write a single high-level requirement for an embedded Hello World. He explains when requirements pay off, how they support testing and scope control, and why you should not write them for every small script. He then lays out five practical rules and applies them to a concrete EHW-001 serial transmission requirement.
Lessons Learned from Embedded Code Reviews (Including Some Surprises)
Jason Sachs recounts a round of motor-controller code reviews and the practical lessons his team learned about quality and tooling. He explains how a simple "ready for review" checklist and automated style checks kept meetings focused on substantive issues, and why choosing the right review tool matters after discovering lost comments in Stash. Read for concrete tips on process, subgit mirroring, vera++, and Upsource.
Introduction to Microcontrollers - Ada - 7 Segments and Catching Errors
Mike demos an Ada implementation of a multiplexed 7-segment driver on the STM32F407 Discovery board, highlighting Ada idioms like protected objects for ISRs and packed-boolean GPIO mapping. The post shows practical timer setup for Timer 6, how to avoid ARR/CNT races, and how Ada's runtime range checks plus a last-chance handler surface out-of-range errors with file and line diagnostics.
Bad Hash Functions and Other Stories: Trapped in a Cage of Irresponsibility and Garden Rakes
A tiny filename decision in MATLAB's publish() can silently swap rendered equations, and Jason Sachs shows why that matters. He reproduces the bug, walks through hash-function basics and collision math, and contrasts safe and unsafe caching strategies. The piece then broadens into practical lessons about software fringes, legacy constraints, and the usability traps that leave engineers repeatedly stumbling over avoidable design choices.
Graphical medicine
Although an appealing user interface is a good option for any device, in medical applications it can be a life saver.
Linear Feedback Shift Registers for the Uninitiated, Part XV: Error Detection and Correction
CRCs and Hamming codes look a lot less magical when you view them as redundancy with a purpose. Jason Sachs walks from parity bits and checksums into finite-field polynomial arithmetic, then shows how CRCs map cleanly onto LFSRs and how Hamming codes use syndromes to locate single-bit errors. It is a practical tour of error detection and correction, with enough worked examples to make the theory feel usable.
nRF5 to nRF Connect SDK migration via DFU over BLE
This writeup contains some notes on how I was able to migrate one of my clients projects based on the nRF5 SDK, to nRF Connect SDK (NCS) based firmware, via a DFU to devices in the field over BLE.
Linear Feedback Shift Registers for the Uninitiated, Part IV: Easy Discrete Logarithms and the Silver-Pohlig-Hellman Algorithm
Discrete logarithms can be either trivial or infeasible depending on how group elements are represented, and Jason Sachs shows a practical route when they are intentionally easy. This article walks through using LFSRs as fast counters, why a smooth group order matters, and how the Silver-Pohlig-Hellman algorithm plus the Chinese Remainder Theorem recovers exponents in GF(2) with small prime factors.
Embedded Development Is Broken. Here's the Strategy I'm Betting My Company On.
Here's a 79-word summary: Embedded software complexity is growing about 4x per decade while developer productivity grows 1.5x, and regulations like the EU CRA are widening the gap further. After running a firmware services company through this shift, I've come to see three things separating the teams that are pulling ahead: using AI where the work is actually hard, designing security in from day one, and reading the standards that govern their market (62304, 26262, CRA) before writing code, not after.
Coding Step 3 - High-Level Requirements
Stephen Friederichs turns the series toward embedded code by showing how to write a single high-level requirement for an embedded Hello World. He explains when requirements pay off, how they support testing and scope control, and why you should not write them for every small script. He then lays out five practical rules and applies them to a concrete EHW-001 serial transmission requirement.
Your Unit Tests Won't Find the Wolves: Why Embedded Developers Should Be Fuzzing
You test the happy paths. You check the well-formatted packets and the expected inputs. But real users don't read manuals, and real data doesn't follow your protocol spec. Fuzzing throws millions of randomized inputs at your code to find the crashes you never thought to look for. Here's why it matters for embedded systems.
Favorite Tools: C++11 std::array
Firmware teams that avoid malloc or new need safer alternatives, and this post makes a strong case for C++11 std::array as that alternative. It highlights zero-overhead, type-safe, compile-time buffers and points to an ESP32 LED-strip demo where NUM_PIXELS_ fixes RAM usage at build time. Read it to see std::array used with std::rotate, passed to C libraries via data(), and as a low-risk path to std::vector later.
Product quality: belief or proof?
Embedded software development is a challenging activity, so it is essential to have tools and IP that is of the best quality. However, assessing that quality can be, in itself, a challenge.
Data Validity in Embedded Systems
Real-world sensors and serial links often deliver garbage, and embedded software must recognize and handle invalid inputs before they cause failures. In this post Stephen Friederichs walks through practical validity checks, from simple range tests and sentinel values to hardware status flags and timing checks for stale data. He also outlines safe responses, from graceful degradation to fail-safe shutdowns, so firmware behaves predictably in the unexpected.
Data Validity in Embedded Systems
Real-world sensors and serial links often deliver garbage, and embedded software must recognize and handle invalid inputs before they cause failures. In this post Stephen Friederichs walks through practical validity checks, from simple range tests and sentinel values to hardware status flags and timing checks for stale data. He also outlines safe responses, from graceful degradation to fail-safe shutdowns, so firmware behaves predictably in the unexpected.
Embedded Software Creation II - European Normative & Legislation
If you are building a product for the European market, standards and directives can determine whether it can legally ship with a CE mark. Maykel Alonso breaks down how European normative and legislation fit together, what “New Approach” and “Global Approach” mean, and which directives commonly affect embedded products. The post also points engineers to the Official Journal of the European Union and a useful directive reference table.
Never use Float or Integer
Ada treats numbers as more than just numbers, and that changes how embedded code fails. This post shows why you should avoid using Float and Integer directly, then demonstrates how distinct types, ranges, and subtypes let the compiler catch unit mix-ups and out-of-range values before runtime. It also shows the same code running on a Raspberry Pi Pico, and briefly introduces SPARK for proving correctness.
Remember Y2K?
There was fear that the turn of the century at the end of 1999 would cause problems with many embedded systems. There is evidence that the same issue may occur in 2038.
Scorchers, Part 4: Burned by the Happy Path (Simon Says)
Designs that only work along the happy path break in real use, causing frustration and sometimes safety risks. Jason M. Sachs uses everyday examples from microwaves to car Auto Park logic to show how mutable software and physical state create brittle behavior. He outlines practical firmware fixes such as clear state machines, sensor or user-driven resynchronization, soft-start delays, and a ‘‘Drunken Happy Path’’ fuzzing approach to find real-world failure modes.
Graphical medicine
Although an appealing user interface is a good option for any device, in medical applications it can be a life saver.
Autonomous vehicle - design questions to ponder
Designing an autonomous or remotely controlled vehicle forces tradeoffs between purpose, environment, and terrain. This post lists practical questions to drive decisions about mission profile, ruggedization, power and data handling, sensing, and modularity so your robot survives and completes tasks. Engineers will get a concise checklist to provoke design choices and avoid common failure modes in real-world deployments.
Product quality: belief or proof?
Embedded software development is a challenging activity, so it is essential to have tools and IP that is of the best quality. However, assessing that quality can be, in itself, a challenge.
Ten Little Algorithms, Part 8: Miller-Rabin Primality Test (and Living with Uncertainty)
Part 8 of the Ten Little Algorithms series: A look at the Miller-Rabin primality test, along with Pollard's rho algorithm for factoring, and some perspectives on very low levels of uncertainty.
Beyond the Packet: Designing Reliable Serial Communication for Embedded Systems
Serial communication between microcontrollers sounds simple until the protocol quietly breaks your system. Prabo Semasinghe walks through the design steps for building a robust communication framework: packet structure, error detection, acknowledgment handling, state machine design, and the failure-mode testing that actually proves it works.
