What do you do if your Robotics project faces setbacks and how can you learn from them?

Robotics project setbacks can teach and improve the project. Stay calm, analyse the root cause, identify lessons learned, adapt and adjust your approach, seek expert advice, break problems down into manageable issues, experiment and iterate to find solutions, document and share learnings, maintain resilience, review and reflect regularly, and stay positive and persistent to overcome these challenges. Simplifying challenges, concentrating on long-term goals, and testing multiple ways can assist find solutions. Documenting and sharing learnings with your team or community helps prevent future failures and promote continual progress.

The first step is to profile the areas where your robot isn’t doing what it should. It is an investment to develop solid metrics that can inform you and pin point the overall bad performance to a precise area of the robot. If you do them well and maintain them, you can refer to them in future to inform what is the right development direction to take. Ideally you should bake metrics and observability from the early life of the product, if you didn’t and you experience problems you need to focus on this first starting by profiling the most urgent pain point. Data gathering does not only help decision making, it can be engineered to enable AI later on in the development cycle

Isolate the problem: Identify whether the issue is mechanical, electrical, or software-related. Examine components: Check the physical condition of the components and replace any that are damaged or worn out. Review code: Look for errors or bugs in your code and test it thoroughly. Run diagnostic tests: Use diagnostic tools to identify any issues with the hardware or software. Document the process: Keep a record of the steps taken to troubleshoot the problem, including any solutions found. Learn from the experience: Use the knowledge gained from analyzing the failure to improve future projects and avoid similar issues.

Usually, failure to not meeting the performance metrics (Throughput, Reliability, Productivity, Lead Time etc.) is the symptom hence it is critical to deep dive and identify the root cause early, through error logs, measurable data, site visit, user feedback etc. Most of the times project can revive through minor design and workflow optimizations. However, sometimes it needs to be rolled back for business continuity. Stay calm and perseverant, document the details of identified root cause/design gap and create a team of experts to brainstorm and find a practical solution with a timeline. Once ready, pilot in a test environment and fine tune. Go-live in stages by monitoring the key metrics closely for continuous improvement till handover.

Working together is very important in robotics, especially when you run into problems. It's helpful to ask for opinions and ideas from people outside of your team because they can bring different perspectives and creative answers. Also, being involved in a larger robotics community gives you the chance to share knowledge and resources, which makes it easier to overcome difficulties. It's essential to keep in mind that teamwork improves the way we solve problems when building complex robotic systems.

Ensure you put your team in a condition that would allow to iterate and fail quickly. If you know what to improve you will want to try out new ideas as soon as you can and quickly find out if they worked or not. Remember, robotics is a new subject and you need to support experimentations and expect trial and errors if you want to be successful. Simulators, continuous deployment, collecting data at every possible opportunities are some ideas to ensure your team can quickly prototype solutions, you don’t want to have several weeks delay between a completed feature and field trials

Project setbacks may come, but we can minimize the chance by developing solid requirements. If those requirements are based on the user needs, and if they are flowed down to sub-system, component, and process requirements / specs, then the system is much less likely to encounter scope creep, cost overruns, and quality issues that cause so many project delays.