ACO-FILL

ACO-FILL is developing an acoustic and AI-based platform technology for non-contact fill-level detection in containers. The method uses the acoustic impulse response of the container as a state-dependent signature: a defined acoustic excitation pulse is introduced into the container; a microphone records the resulting impulse response, from which signal-processing and machine-learning methods extract robust features and derive fill-level classes and trends.

The project addresses the technological gap between low-cost but unreliable approaches, such as simple switches, and established sensing methods, including ultrasound, radar, capacitive and gravimetric sensing. In certain applications, these established methods are limited by cost, installation effort, maintenance requirements or contact with the medium. A European patent application (EP4628854 A1) protects the innovative approach.

The project starts at TRL 3-4, with the underlying physical principle demonstrated and a prototype measurement setup in place. Its objective is to develop an integrated demonstrator (TRL 6) that performs measurements autonomously and processes data on-device or at the edge. This includes a federated learning architecture for continuous model improvement based on distributed field data. Standardised interfaces will enable integration into higher-level systems. Smart waste management, using bins and containers, will serve as the pilot and initial application to demonstrate transferability to other container applications and industries.

ACO-FILL strengthens the research profile of Trier University of Applied Sciences - Environmental Campus Birkenfeld at the intersection of acoustic sensing, embedded systems and data-driven modelling, while laying the foundation for a commercially viable and broadly scalable platform technology.