This work presents concepts of the use of algorithms inspired by the functions and properties of the nervous system in dense wireless networks. In particular, selected features of the brain consisting of a large number of nerve connections were analyzed, which is why they are a good model for a dense network. In addition, the action of a selected cells from the nervous system (such as neuron, microglia or astrocyte) as well as phenomena observed in it (e.g. neuroplasticity) are presented.

Non-orthogonal multiple access (NOMA) has received tremendous attention for the development of 5G and beyond wireless networks. Power-domain NOMA works on the concept of assigning varying power levels to users within the same frequency and time block. In this paper we propose a novel power allocation approach that uses the Zipf distribution law that satisfies the basic condition of a NOMA system. The Zipf PA is characterized by the simplicity and ease of implementation that allows to extend the capacity of the system to support a large number of users. The numerical results show that the system achieves high throughput and energy efficiency without any parameter optimization constraints as well as improved capacity by increasing the number of users compared to the NOMA system with existing power allocation techniques.

One of the ways to improve calculations related to determining the position of a node in the IoT measurement system is to use artificial neural networks (ANN) to calculate coordinates. The method described in the article is based on the measurement of the RSSI (Received Signal Strength Indicator), which value is then processed by the neural network. Hence, the proposed system works in two stages. In the first stage, RSSI coefficient samples are taken, and then the node location is determined on an ongoing basis. Coordinates anchor nodes (i.e. sensors with fixed and previously known positions) and the matrix of RSSI coefficients are used in the learning process of the neural network. Then the RSSI matrix determined for the system in which the nodes with unknown positions are located is fed into the neural network inputs. The result of the work is a system and algorithm that allows determining the location of the object without processing data separately in nodes with low computational performance.

Data security is one of the prime concerns in wireless networks. PLKG has been emerging as an attractive alternative to traditional cryptographic techniques. PLKG is more computationally efficient than cryptography. Moreover, PLKG using Principal component analysis (PCA) as pre-processing may further save computations. This paper proposes three mechanisms to select components of PCA which are based on Information content, Mean and Histfit. Bit Disagreement Rate (BDR) is compared for each mechanism. Histfit based method is found to be best. Since only two components are supposed to be processed for key generation, it is computationally efficient/ power efficient too.

A key role in the development of smart Internet of Things (IoT) solutions is played by wireless communication technologies, especially LPWAN (Low-Power Wide-Area Network), which are becoming increasingly popular due to their advantages: long range, low power consumption and the ability to connect multiple edge devices. However, in addition to the advantages of communication and low power consumption, the security of transmitted data is also important. End devices very often have a small amount of memory, which makes it impossible to implement advanced cryptographic algorithms on them. The article analyzes the advantages and disadvantages of solutions based on LPWAN communication and reviews platforms for IoT device communication in the LoRaWAN (LoRa Wide Area Network) standard in terms of configuration complexity. It describes how to configure an experimental LPWAN system being built at the Department of Computer Science and Telecommunications at Poznan University of Technology for research related to smart buildings.