The increasing demand for higher-quality specialty coffee is making the industry pay special attention to temperature management during roasting. Instead of keeping the temperature at an exact point, managing the temperature over time is more important, directly impacting flavor development on the final roasted beans. Generally, the temperature within the coffee roaster is managed by a human operator, making it hard to recreate a similar temperature profile between different roasting batches. This paper proposes a novel embedded system that implements trajectory tracking control based on Adaptive PID-M with the goal of eliminating over-reliance on an operator's skill to manage the repetitive roasting process. Experimental results show that the implemented system is able to recreate the temperature profile with a mean square error (MSE) of 4.56 C which is a huge improvement from human operators with an MSE of
21.7 C. Implementing this system will allow coffee roasting shops to improve quality control and reduce waste from inconsistent roasting batches
