Indoor positioning systems (IPS) have great potential to define the location of objects with no GPS or other radionavigation data. Such systems include location estimation algorithms based on time series data from Wi-Fi, BLE, and other devices. Algorithms use to tracking location in real-time and obtain a trajectory close to the actual path. This, in turn, opens up opportunities for finding typical pathways, queues, and bottlenecks in various indoor places. IPS are often used in healthcare, and they are an essential part of the organization of internal processes in the case of a virtual hospital. In this research, we use iBeacons equipment because of its low cost and ease of use. However, the signals received at the objects have high noise, and the location estimation algorithms have an error that accumulates over time. This paper considered two ways to solve high noise: a probabilistic-based method and a neural network method. These algorithms have closer errors (2.11 - 0.96 m), but using the neural network method makes it possible to increase the performance of the indoor positioning algorithms.