Glucose sensors have improved and facilitated therapy for type 1 diabetes. However, they are still not capable to sense all physiological signals and to act in a closed-loop. Pancreatic β-cells have been shaped during evolution as biological sensors and offer the advantage to integrate all physiological signals in addition to glucose. Moreover, biosensors based on these cells may also serve for non-invasive and continuous long-term characterization of β-cells, drug research, tissue engineering and pre-transplantation quality control. β-cells alter their electrical activity upon exposure to glucose and physiological hormones and we have used these properties to design a biosensor. To this end signals were recorded extracellularly from islet cells kept on multi-electrode arrays. Slow and rapid oscillations were observed, both modulated by glucose. Especially slow oscillations are very robust and have an excellent signal/noise ratio. Signal processing functions were designed to separate the two activities to extract and analyze relevant parameters. These parameters correlate very well with either increasing or decreasing glucose concentrations. An electronic device is under construction, based on an embedded FPGA capable of processing multiple channels in parallel. In the future, such a device shall be used as a portable real-time biosensor regulating insulin delivery from a pump.