Injection molding is a highly dynamic process with the need to be controlled if highly accurate technical parts are to be repeatedly produced on a mass production scale. The need to control the process evokes the necessity for sensors enabling detection of the current process conditions. This work presents the analysis, implementation and investigation of a wireless in-mold sensor, called the acoustic-emission sensor. The sensor makes use of structure-borne sound as transmission medium eliminating the need of undesirable wiring connecting the sensor head to the exterior surface of the mold. The sensor detects the melt front location at (multiple) predetermined positions. A movable pin, e.g. an ejector pin, is accelerated through the melt pressure within the passing melt front and the pin impacts on a resonant structure. The structure is excited and oscillates at its resonant frequency and distributes the oscillations in form of structure-borne sound within the whole metal mass of the mold. An accelerometer on the outside surface of the mold enables the detection and further recognition of the resonant structure oscillation. Consequently the temporal position of the melt front can be measured at predetermined positions. Simultaneous detection at multiple locations is possible. This is achieved by designing individual resonators to have distinct resonant characteristic. Signal processing is used to separate the sounds emanating from the resonators. In addition to two classical approaches a new linear algebraic approach is introduced, the frequency pattern recognition method. The new algebraic enables a least squares approximation for the instant of time when the resonator was excited, i.e., when the melt front reached the sensor. The new method uses discrete orthogonal polynomials and constrained basis functions. Additionally, the method yields the complete covariance propagation, from which an uncertainty can be computed via the inverse student-t distribution. In this manner, the time of excitement and a confidence interval can be determined. This is of special interest in instrumentation. Experimental results verify the good performance of the acoustic-emission sensor concept when implemented in an injection mold. In comparison to conventional in-mold sensors the acoustic-emission sensor shows at least identical results in terms of response time which is a very important value for melt front position detection. The acoustic-emission sensor was tested in combination with the frequency pattern recognition method in a long term test showing reliable automatic melt front detection with single or multiple implemented resonators.