Modelling the mmWave Channel Based on Intelligent Ray Launching Model Jialai Weng 1 , Xiaoming Tu 2 , Zhihua Lai 3 , Sana Salous 4 Jie Zhang 5 1 Department of EEE, The University of Sheffield, Sheffield, S1 3JD, UK, jialai.weng@sheffield.ac.uk 2 Department of EEE, The University of Sheffield, Sheffield, S1 3JD, UK, xiaoming.tu@sheffield.ac.uk 3 Ranplan Wireless Network Design Ltd, Sheffield, S1 4DP, UK, zhihua.lai@ranplan.co.uk 4 School of Engineering and Computing Sciences, Durham University, DH1 3LE, UK, sana.salous@durham.ac.uk 5 Department of EEE, The University of Sheffield, Sheffield, S1 3JD, UK, jie.zhang@sheffield.ac.uk Abstract—Path loss modelling and prediction is essential in indoor network planning. To find an efficient and accurate path loss model benefits the network planning and optimisation process. In this work, we propose an efficient multipath empirical path loss prediction model based on an intelligent ray launching algorithm for 60GHz indoor channel. The comparison between the simulation result and measurement shows good agreement. This demonstrates that the proposed model is suitable for indoor network planning application. Index Terms—antenna, propagation, measurement. I. I NTRODUCTION Wireless network planning and optimisation is a key com- ponent in network deployment. To meet the demand of optimal network service with minimal resource and cost, network planning plays a key role in the deployment of the networks. Channel map is one of the key tools that are essential in the planning and optimisation of the networks geographically. The channel map is to indicate the channel characteristics over the geographic map of the network deployment space. The channel map is used as an essential tool to describe the channel condition geographically and then facilitates the network deployment in the design space. Channel maps are widely constructed through computer simulation based electromagnetic propagation predication tools. One of the popular model to simulate the electromag- netic propagation is based on ray tracing. The early work [1] first introduced the simulation base channel predication soft- ware tools Wise. It is based on the electromagnetic propagation principles of geometrical optics (GO) and Uniform Theory of Diffraction (UTD). The software CINDOOR [2] is based on the same physical propagation simulation and is specifically designed for indoor environment. The work [3] extended the work to include diffusive diffraction in ray tracing. More recent work [4] also applied the ray tracing simulation to construct MIMO channels. Another method is based on the numerical solution of the wave propagation equation and the Maxwell equations. The finite-difference time domain (FDTD) method gives a numerical solution to the electromagnetic field in time domain. Following the FDTD algorithm, the work [5] proposed a finite element based numerical method to build channel map efficiently. The millimetre wave (mmWave) channel is proposed as a promising technique for the future wireless networks. The channel has many unique characteristics. To deploy networks equipped with mm wave channels, a channel map construction tool that is capable of building mmWave channel map is essential. However, such a tool is still missing. In this work, we proposed a simulation based channel map construction tool based on an intelligent ray launching algorithm (IRLA) [6], especially for modelling the path loss value of 60GHz indoor channel. It exploits the characteristics of the mmWave chan- nels to simplify the computational complexity of the traditional simulation tools, while still preserves high degree of accuracy of the original physical electromagnetic propagation models. The simulation model is then verified by the measurement result in a real propagation environment. The rest of the paper is organised as follows. Section II described the new ray tracing based model. Section IV is the comparison and analysis of simulation. Section V is the conclusion. II. THE PATH LOSS MODEL BASED ON RAY LAUNCHING A widely used path loss model is the empirical model of log distance model. The model is simply given as pl(d)= PL(d 0 ) + 10γ log 10 ( d d 0 ) (1) where pl(d) is the path loss value at the distance d; PL(d 0 ) is the path loss value at the reference distance d 0 ; γ is the path loss exponent. The path between the transmitter and receiver can be seen as the solely propagation path chosen in the model. Hence, the model can be seen as a single propagation path model. It is an empirical model widely used in modelling the path loss of propagation channel [7]. Wireless channel is widely modelled as multipath propaga- tion. In multipath propagation model, the channel is written as H(d)= n  1 a(d) i exp(-jθ(d) i ) (2) where H(d) is the channel impulse response at distance d; i is the multipath index; n is the number of the multipath components; a i and θ i are the amplitude and phase of the i-th multipath component, respectively. For calculating the