Development of a Smart Application for Indoor Navigation (INMaps)

Indoor navigation framework plays a significant role in day today life. In the outdoor application, GPS signals identifies the exact location or points of the user, but it is not true in indoor scenarios. To overcome this issue there are various techniques have been devised to predict the position and identify the location of the user in the indoor application. These include attaching sensors to the shoes or suites or ceiling of the building and using WIFI signals to predict the position and navigate user to the destination. But not all the time this is possible as the sensors may get damaged and signal interferences may lead to large accumulated errors in the results. This paper proposes a novel technique to predict the position and navigate the user making use of the sensors present in the smartphone. Use of smartphone eliminates the need for external sensors to be attached to the shoes or suits, as it has inbuilt magnetometer, accelerometer, and gyroscope. The data generated by these sensors are used to estimate the pitch and roll values and also the heading information. The step count and the time at which the user performs each step are calculated using peak detection algorithm. A rule-based algorithm is proposed to estimate the step length and the smart phone beacon signals are used to provide the heading information. A voice based guiding facility is also built in to alert the user in case he selects the wrong path. With all these features, the proposed system certainly helps the user to navigate correctly from source location to the intended destination. The experiments are carried out in different scenarios under various realistic conditions and the results displays that, the proposed method achieves a high position accuracy with significant reduction in the error (less than 2.5%) and performs well compared to the conventional estimation methods. The performance is assessed in terms of displacement and root mean square error and compared with the position-estimation method (Poulose in IEEE Access 7: 11165–11177, 2019).