Channel Coding in Molecular Communication

This dissertation establishes and analyzes a complete molecular transmission system from a communication engineering perspective. Its focus is on diffusion-based molecular communication in an unbounded three-dimensional fluid medium. As a basis for the investigation of transmission algorithms, an equivalent discrete-time channel model (EDTCM) is developed and the characterization of the channel is described by an analytical derivation, a random walk based simulation, a trained artificial neural network (ANN), and a proof of concept testbed setup. The investigated transmission algorithms cover modulation schemes at the transmitter side, as well as channel equalizers and detectors at the receiver side. In addition to the evaluation of state-of-the-art techniques and the introduction of orthogonal frequency-division multiplexing (OFDM), the novel variable concentration shift keying (VCSK) modulation adapted to the diffusion-based transmission channel, the lowcomplex adaptive threshold detector (ATD) working without explicit channel knowledge, the low-complex soft-output piecewise linear detector (PLD), and the optimal a posteriori probability (APP) detector are of particular importance and treated. To improve the error-prone information transmission, block codes, convolutional codes, line codes, spreading codes and spatial codes are investigated. The analysis is carried out under various approaches of normalization and gains or losses compared to the uncoded transmission are highlighted. In addition to state-of-the-art forward error correction (FEC) codes, novel line codes adapted to the error statistics of the diffusion-based channel are proposed. Moreover, the turbo principle is introduced into the field of molecular communication, where extrinsic information is exchanged iteratively between detector and decoder. By means of an extrinsic information transfer (EXIT) chart analysis, the potential of the iterative processing is shown and the communication channel capacity is computed, which represents the theoretical performance limit for the system under investigation. In addition, the construction of an irregular convolutional code (IRCC) using the EXIT chart is presented and its performance capability is demonstrated. For the evaluation of all considered transmission algorithms the bit error rate (BER) performance is chosen. The BER is determined by means of Monte Carlo simulations and for some algorithms by theoretical derivation.