This paper discusses the merits of Mechanomyography (MMG) sensors in capturing and isolating muscle activity in high interference environs, with application to `hands free' gear shifting on a bicycle for users with limited extremity movement. MMG (acoustic) muscle sensing provides a simple and rugged alternative to physiological sensing for machine interface in the field, but suffers from interfering artifacts (in particular motion) which has limited its mainstream use. We introduce a system fusing MMG with a filter based on Inertial Measurement (IMU) to isolate muscle activity in the presence of interfering motion and vibrations. The system identifies user-initiated muscle trigger profiles during laboratory testing, allowing parameterization of MMG and IMU signals to identify purposeful muscle contractions (triggers) and to omit false triggers resulting from cycle/road vibration or rider movement. During laboratory testing the success rate of trigger identification was 88.5% while cycling with an average of 0.87 false triggers /min. During road testing the success rate was 72.5% and false triggers were more frequent at 3.7 /min. These results hold strong promise for alternative triggering mechanisms to the standard bar-end shifters used in current off-the-shelf cycling group sets, enabling amputees or people of reduced arm or hand dexterity to change gears while riding. Further testing will explore the use of signal filters on MMG data and further use of IMU data as feedback to increase false triggers rejection. Wider applications include a broad range of machine-interaction research.