In this work, we study a wireless communication system designed to handle multi users, each communicating with their own frequency bands. We analyse the performance of the system - its information capacity – while considering different number of users, and different configurations for the multipath propagation scenarios for each user and their base frequency. Our interest is to discover worst- (blind spots) and best-case scenarios (maximal information transmission). Our results show that a prior effective choice of parameters, such as the natural frequency of the chaotic signal, lead to significant throughput improvement and, in some cases, even enable an error-free communication. It is often to be expected that physical constrains in the channel may block higher-frequency signals used in standard non-chaotic communication systems. For our chaos-based communication system, we show that a significant wireless channel constraint, the multipath propagation, becomes less disruptive, the higher the frequency of the user. That is a win-win result. Not only does the higher frequency allow more information to be transmitted, but it also prevents multipath interference. Moreover, there is an ideal relationship between the time of propagation for the signals with the user natural frequency that results in a communication with no interference due to multipath, thus effectively creating a wireless communication system with similar gains as that obtained for wired communication where multipath propagation is not a significant issue. As an application of our system, we study a sensor network configuration where sensors have the same frequency and show that given a particular set of time propagation for the signal to leave the sensor and to arrive in the receiving station, we can determine an optimal natural frequency for the sensors to maximise information transmission. Alternatively, the contrary is also true, and that is that optimal natural frequencies can be deduced, given the time propagation configurations. Some optimal spatial configurations with no interference due to multipath will be deduced from this analysis.