IndoorGuide Pedestrian Navigation based on a Foot-Mounted IMU

Since the beginning of the 21st century, personal indoor navigation and localization systems are an important research area for engineers and scientists in the field of navigation. Public funding from the government and global operating companies are the driving forces on this wide and cost-intensive research area. The focus of this research area is to get robust and precise localization techniques for indoor environments. An enhanced personal navigation system handles transitions between indoor and outdoor environments by applying intelligent data fusing techniques. The key is to get a robust navigation solution for indoor and outdoor scenarios. Combining an intelligent data fusing strategy with multi sensor system should yield a navigation system, which is able to support the user with reliable and precise navigational information. Mobile personal navigation systems provide assistance in several applications. Especially during search and rescue (SAR) or security missions, a personal navigation system can enhance the safety of the personnel. Finally, consumers or visually impaired persons can profit from a personal navigation system for indoor and outdoor environments. All user groups can be assisted - not to lose the orientation - in complex and previously unknown buildings. In this publication, we will present a personal navigation system based on a foot-mountable unit for precise and robust indoor positioning without any other aiding sensor. Core of the foot unit are two ultra-low-cost, microelectromechanical system (MEMS) based IMUs. Both MEMS sensors are 9-axis inertial sensors to measure the acceleration, the angular rate and the local magnetic field. All integrated components are ultra-low-priced products from the mass market. The size of the foot-mountable unit is very small. The housing of the system has a dimension of 4.2cm x 4.8cm x 1.5cm (1.65in x 1.89in x 0.59in), including a battery pack, a Bluetooth module and an ARM Cortex-M3 microcontroller. The foot unit can be connected wirelessly or via USB-cable to a tablet computer or a smartphone. The presented prototype of our new system was designed and developed to make personal navigation available for the consumers. As reference system, we use an extended foot-mountable unit, which was developed at our institute. This module is equipped with a further and more precise MEMS-based IMU, the ADIS16448 from Analog Devices. This inertial sensor has a higher sensor grade than the ultra-low-cost MEMS IMUs MPU-9250 from InvenSense used in the new foot unit. Furthermore, the reference system is equipped with a GNSS module for absolute positioning outdoors and a barometric pressure sensor for altitude measurements. The GNSS module inside the reference foot unit is not necessary, if the tablet computer or the smartphone can provide an initial position or position updates (after minutes) with sufficient accuracy. Current GNSS technologies are not reasonably effective for indoor positioning, due to multi-path effects, a limited availability and a low signal-to-noise ratio. Due to challenging operating conditions inside buildings, our strategy is reduce the drift of the inertial sensor of our new foot unit with our already published Finite State Machine (FSM) based step detection technique for pedestrian navigation systems. This approach is founded on medical research findings of the human gait and analyze the typical acceleration and angular rate waveforms during a human gait cycle. The state machine maps different motion phases of the human gait in chronological order onto a graph. This approach analyses the motion of the foot and allows the navigation filter to accurately apply zero velocity updates (ZUPTs) at the right time. Due to the precision and robustness of our FSM-based technique we could save costs by substituting additional aiding sensors. Therefore, we designed and released the developed foot unit without an air pressure sensor and GNSS module. The successful operation of the new foot-mountable system for pedestrian navigation is validated in field tests and a comprehensive demonstration of the system will be given.