Due to its several implementation advantages, orthogonal frequency division multiplexing (OFDM) has become the most frequently used waveform in modern wireless communication technologies. However, OFDM has some vulnerabilities that may limit its advantages. In order to alleviate OFDM's vulnerabilities, several waveforms have been proposed within the context of 5G research activities. Universal-filtered OFDM (UF-OFDM) is one of the most effective waveforms due to its OFDM-like structure, easier implementation procedure and robustness against synchronization errors. Most of the recent studies on novel waveforms generally target filter optimization or performance improving techniques. In this paper, as a novel approach, UF-OFDM with filter shift keying (OFDM-FSK) is proposed. Here, we consider the filter realizations to carry information through index modulation mechanism. Accordingly, utilized filter of each UF-OFDM subband is selected from a filter set based on the incoming data bits, and an increase in data rate is obtained. At the receiver side, accumulated maximum-likelihood decision metric in each subband is computed. The filter realization, as well as data symbols, that provide the minimum decision metric in each subband are detected. Hence the throughput of the conventional waveform is increased in an efficient and practical manner. The performance of OFDM-FSK system is evaluated through extensive computer simulations. To assess its real-time performance, real-time experiments are also carried out by using software defined radio (SDR) nodes. In the created testbed, a wireless synchronization structure is provided and the considered waveforms are tested in a setup with imperfect synchronization. This experimentation provides realistic insights about the benefits of OFDM-FSK. As it will be demonstrated via both computer simulation and test results, OFDM-FSK provides significant throughput and error performance benefits over UF-OFDM.