Modeling and Evaluation of Molecular Communication Systems

Synthetic molecular communication (MC) is an emerging research field that employs signaling molecules as information carriers, inspired by natural processes. Biological examples, such as hormone signaling through the bloodstream and synaptic transmission in the nervous system, demonstrate the efficiency and versatility of molecular information transfer. Translating these mechanisms to engineered systems offers promising solutions for micro- and nanoscale applications where conventional electromagnetic (EM)-based communication faces fundamental limitations, particularly in biomedical contexts. Potential applications include smart health monitoring, targeted drug delivery, intra-body sensor networks, and environmental monitoring. Nevertheless, significant challenges must be addressed before practical implementations can be achieved. In this dissertation, some of these challenges are addressed by modeling and evaluating novel aspects of MC systems. Specifically, a new modulation scheme and transmitter (TX) design is proposed, which are experimentally validated within the first closed-loop MC testbed. In addition, new performance metrics are introduced to support systematic evaluation and benchmarking of multi-link MC systems. The main contributions of the dissertation can be summarized as follows. Resource-Efficient TX Model for MC: Conventional MC systems typically assume that a TX stores, releases, and possibly replenishes signaling molecules; however, storing large quantities of molecules and ensuring replenishment are often impractical. To overcome these challenges, media modulation is proposed, where the TX encodes information by modulating the state of switchable signaling molecules already present in the medium. In particular, the use of photochromic molecules that undergo reversible, light-induced state transitions are investigated as candidates for media modulation-based MC systems. Through comprehensive modeling, it is demonstrated that media modulation can, in theory, achieve reliable transmission and represents a promising alternative to conventional molecule-emitting MC systems. First Long-Term Closed-Loop Experimental Setup for MC: To demonstrate the practical feasibility of media modulation, a fluid-based experimental MC testbed inspired by the cardiovascular system, operating within a closed-loop tube network, is presented. As switchable signaling molecule, the photochromic green fluorescent protein variant Dreiklang (GFPD) is employed, which can be reversibly switched between fluorescent and non-fluorescent states using light triggers. This property enables repeated writing, erasing, and reading of information with the same GFPD molecules, thereby eliminating the need for a molecule-emitting TX. In long-term experiments lasting up to 250 minutes, reliable, error-free transmission of 500 bits at a rate of 6 bit/min using a low-complexity detector is demonstrated. ARE and ARTE for Multi-Link MC Systems: A multi-link diffusion-based MC system is studied to gain a fundamental understanding of the benefits of exploiting the spatial dimension in MC systems. In particular, the following research question is addressed: Given specific transmit and receive areas accommodating multiple spatially distributed point-to-point non-cooperative TX–RX links, how densely should they be deployed to maximize the information transmission rate? This problem involves a fundamental tradeoff: Increasing link density enhances spatial reuse, but overly dense deployments induce severe inter-link interference, ultimately reducing the overall transmission rate. To address this, two new performance metrics for multi-link MC systems are introduced, namely area rate efficiency (ARE) and area–time rate efficiency (ARTE). Based on these metrics, it is shown that an optimal transmission link density exists. Semantic Information in MC Systems: Most MC research adopts Shannon’s syntactic notion of information, yet many applications require task-oriented measures. In this work, the semantic information framework of Kolchinsky and Wolpert is applied to bacterial chemotaxis. Using agent-based modeling, it is quantified how much information a chemotactic bacterium extracts from its environment to adapt and survive. The results show that the effective use of information depends strongly on environmental conditions, thereby directly linking communication to survival. This perspective opens new directions for MC design, where system effectiveness is evaluated in terms of task performance rather than information throughput alone.