Sensory feedback information is essential for fish to explore and inhabit various habitats or survive predator-prey encounters. Inspired by fish lateral line sensory systems, numerous flow sensors have been successfully developed based on MEMS technology which has the potential for applications on autonomous underwater vehicles. However, many challenges, such as robustness against extreme underwater conditions, installation, and signal readout, need to be overcome to realize this promise. To overcome the challenges, we present Vibot, the first non-invasive inertial measurement unit (IMU) method to understand and recognize fluid interactions with the robotic fish. Vibot works by using a commercial off-the-shelf IMU to study interactions between a tuna-inspired robotic platform, the Tunabot, and a variety of flow conditions including laminar flow, a Kármán vortex street, and wall effects. We propose a vibration model of fluid Interactions and utilize an Artificial Neural Network (ANN) to analyze and recognize different water flow conditions. Experiments are carried out in a water channel with the Tunabot platform. Results will be presented and compared with previous results using MEMS flow sensors.