Millimeter-wave (mmWave) communication is one of the emerging key components for next generation wireless networks. It exploits unlicensed band at 60 GHz in an indoor environment. Since the wave propagation at mmWave frequencies exhibits a quasi-optical behaviour, the inherent line of sight (LoS) channel models provide new research directions. The major challenges in mmWave communication are higher free space path loss (FSPL) due to its smaller wavelength, quasi-optical behaviour of wave propagation and blockage and penetration losses in indoor scenario. The losses can be compensated using multiple antenna (MA) systems and the sparse channel behaviour of mmWave propagation is compensated by optimal and aligned antenna geometry. However, MA system increases hardware constraints as the number of radio frequency (RF) chains increases. In the recent years, space shift keying (SSK) modulation scheme has been identified as a promising MA technique with single RF hardware requirement. In SSK, only one antenna is activated for transmission at any time instant and the index of this active antenna is used to convey information. Reliable communication between the nodes in a network can be established with the aid of relay node in mmWave communication. In this paper, a bidirectional relay network model using mmWave communication in indoor-LoS environment is considered. The concept of physical layer network coding (PLNC) is applied at the decode-and-forward (DF) relay node to enable the network operates in bidirectional half duplex mode. Specifically, the transceiver at the nodes sends the information symbol by employing SSK modulation and detects the symbol based on the maximum likelihood (ML) detection criterion. By considering optimal antenna geometry operating conditions in LoS environment, the end-to-end bit error rate (BER) probability of the proposed PLNC based bidirectional relay network using SSK in mmWave communication is derived in closed form. Analytical frameworks and findings are also substantiated via Monte Carlo simulations.