Pedestrian Positioning Research Articles

Pose-Invariant Inertial Odometry for Pedestrian Localization

Inertial sensors in smartphones enable the relative state measurements for pedestrian localization without global positioning system (GPS) signals or beacons. Any such navigation method should notice the reality that the phone is placed at uncontrolled variant poses with respect to the human body. In this work, we focus on pedestrian position estimation from the raw inertial measurement unit (IMU) measurements without any constraint on the device carrying manners and propose a novel deep inertial odometry solution. By presenting the expression of continuous rotating, we are able to release the reliance on the unreliable orientation provided by the sensor application program interface (API). Moreover, we propose a novel loss formulation by representing the velocity as the average velocity magnitude and the moving direction. The proposed approach was assessed via the public RoNIN dataset. The localization performance of the public dataset trained network model was then evaluated by real scenario trials in the CUHK campus. Experimental results show that our model is capable of providing robust velocity estimates and generating more accurate trajectories than state-of-the-art inertial odometry methods. Specifically, the localization evaluation in the CUHK campus contains 60-min inertial signals with a length of 3 km. The trained odometry network is with the 30th percentile accuracy of 2.26 m and the 50th percentile accuracy of 4.98 m.

The Improved 3D Pedestrian Positioning System Based on Foot-Mounted Inertial Sensor

The foot-mounted navigation system based on Micro Inertial Measurement Unit (MIMU) can estimate the position of pedestrians. Zero Velocity Update (ZUPT) provides a solution for reducing cumulative positioning errors caused by the integral calculation of the inertial sensor. However, the low accuracy and high noise of the MIMU will greatly affect the measurement accuracy of zero velocity detector, which will lead to poor 3D pedestrian positioning accuracy. An improved generalized likelihood ratio test (GLRT) method is proposed by establishing adaptive thresholds at different gait frequency. The core is to dynamically adjust the detection threshold according to different motion states, so as to determine the constraint relation for detection. In addition, the drift of sensors and the accumulated error of position integration affect the accuracy of attitude and position. Therefore, an improved extended Kalman particle filter (EKPF) is proposed. The improved EKPF constructs the position constraint model in altitude direction by introducing barometer information. The new particle filter distribution function is constructed by constrained multi-source observation information, which makes the posterior probability distribution closer to the real distribution. In order to verify the feasibility and effectiveness of this method, the inertial pedestrian navigation system is established. The experimental results demonstrate that the improved GLRT method performs well in detecting the zero velocity point of the foot, and improved EKPF method reduces the positioning error by 69.25% and 53.57% compared with the EKF and PF, respectively. Therefore, the proposed method can improved the 3D pedestrian positioning accuracy.

Vehicle Pedestrian Detection Method Based on Spatial Pyramid Pooling and Attention Mechanism

Object detection for vehicles and pedestrians is extremely difficult to achieve in autopilot applications for the Internet of vehicles, and it is a task that requires the ability to locate and identify smaller targets even in complex environments. This paper proposes a single-stage object detection network (YOLOv3-promote) for the detection of vehicles and pedestrians in complex environments in cities, which improves on the traditional You Only Look Once version 3 (YOLOv3). First, spatial pyramid pooling is used to fuse local and global features in an image to better enrich the expression ability of the feature map and to more effectively detect targets with large size differences in the image; second, an attention mechanism is added to the feature map to weight each channel, thereby enhancing key features and removing redundant features, which allows for strengthening the ability of the feature network to discriminate between target objects and backgrounds; lastly, the anchor box derived from the K-means clustering algorithm is fitted to the final prediction box to complete the positioning and identification of target vehicles and pedestrians. The experimental results show that the proposed method achieved 91.4 mAP (mean average precision), 83.2 F1 score, and 43.7 frames per second (FPS) on the KITTI (Karlsruhe Institute of Technology and Toyota Technological Institute) dataset, and the detection performance was superior to the conventional YOLOv3 algorithm in terms of both accuracy and speed.

Multiple-Joint Pedestrian Tracking Using Periodic Models.

Estimating accurate positions of multiple pedestrians is a critical task in robotics and autonomous cars. We propose a tracker based on typical human motion patterns to track multiple pedestrians. This paper assumes that the legs’ reflection and extension angles are approximately changing periodically during human motion. A Fourier series is fitted in order to describe the moving, such as describing the position and velocity of the hip, knee, and ankle. Our tracker receives the position of the ankle, knee, and hip as measurements. As a proof of concept, we compare our tracker with state-of-the-art methods. The proposed models have been validated by experimental data, the Human Gait Database (HuGaDB), and the Karlsruhe Institute of Technology and Toyota Technological Institute (KITTI) tracking benchmark. The results indicate that our tracker is able to estimate the reflection and extension angles with a precision of 90.97%. Moreover, the comparison shows that the tracking precision increases up to 1.3% with the proposed tracker when compared to a constant velocity based tracker.

Generation mechanism and simulation research of zipper phenomenon of pedestrian flow in corridor

In this study, the unidirectional pedestrian flow in the corridor is taken as a research object, the generation mechanism of the pedestrian zipper phenomenon is analyzed, and a velocity correction model based on the Voronoi diagram is established for the simulation research. First, the generation mechanism of the pedestrian zipper phenomenon is analyzed from the perspective of optimal visual field and walking comfort of pedestrians. Then the visual attention and visual occlusion of pedestrians are used to describe the factors which affect the zipper deviation during pedestrian movement, the local density of pedestrians is used to describe the walking comfort of pedestrians, the zipper sensitivity coefficient is adopted to describe the willingness of pedestrians to move objectively, and the mechanism of lateral deviation of a single pedestrian is considered to obtain the optimal deviation position of pedestrians. Besides, the Voronoi diagram is introduced to effectively determine the pedestrians surrounding the target pedestrian within the visual field. And the influence of surrounding pedestrians with different distances and directions on the moving velocity of the target pedestrian based on the Voronoi diagram is considered. Then, a velocity correction model of pedestrians based on the Voronoi diagram is constructed, whether the pedestrian has a subjective willingness to deviate is considered, and the deviation rule is embedded to simulate and reproduce the zipper phenomenon of pedestrians. The simulation results truly reproduce the normal pedestrian flow through the corridor and show that our model can overcome the deficiency of the jitter and overlap phenomenon of the traditional social force model. The self-organized pedestrian flow with uniform distribution and the pedestrian zipper effect can also be observed. Furthermore, through the simulation results, we can see that the number of zipper layers for pedestrians is proportional to the width of the corridor. The comparison of simulated pedestrian data with the empirical data indicates that the fundamental diagram of velocity-density relation of our model is in good agreement with the empirical data. A comparison between with and without considering the zipper effect shows that the larger the proportion of pedestrians actively willing to laterally deviate, the more helpful it will be to improve the moving velocity, comfort and space utilization of pedestrians in the corridor.

UWB/INS Integrated Pedestrian Positioning for Robust Indoor Environments

Goal: Accurate robust ultra wideband (UWB) pedestrian indoor positioning becomes a challenging task when faced with robust indoor environments such as non-line of sight (NLOS) due to refraction of signals, multipath effect, etc. When the UWB system is faced with such a robust environment, its positioning accuracy is difficult to reach the centimeter level and the reliability of the system operation needs to be improved. In order to solve such problems, an ultra wideband/inertial navigation system (UWB/INS) integrated pedestrian navigation algorithm is proposed. On the one hand, UWB-based joint state estimation particle filtering is used for position calculation, and on the other hand, INS-based zero-speed update (ZUPT) algorithm is used for navigation information solution. Under the framework of the INS error equation, the navigation information fusion of the two systems is carried out. In the simple pedestrian environment, the average positioning accuracy of the UWB/INS algorithm is 53.8% higher than that of the UWB algorithm, 40% higher than the INS algorithm and 31% higher than Original UWB/INS algorithm. Under the robust pedestrian indoor positioning, the average positioning accuracy of the UWB/INS algorithm is 39.7% higher than that of the UWB algorithm, 37.5% higher than the INS algorithm and 53% higher than Original UWB/INS algorithm. The results of two sets of pedestrian indoor positioning experiments demonstrate the effectiveness of our approach.

HPIPS: A High-Precision Indoor Pedestrian Positioning System Fusing WiFi-RTT, MEMS, and Map Information.

In order to solve the problem of pedestrian positioning in the indoor environment, this paper proposes a high-precision indoor pedestrian positioning system (HPIPS) based on smart phones. First of all, in view of the non-line-of-sight and multipath problems faced by the radio-signal-based indoor positioning technology, a method of using deep convolutional neural networks to learn the nonlinear mapping relationship between indoor spatial position and Wi-Fi RTT (round-trip time) ranging information is proposed. When constructing the training dataset, a fingerprint grayscale image construction method combined with specific AP (Access Point) positions was designed, and the representative physical space features were extracted by multi-layer convolution for pedestrian position prediction. The proposed positioning model has higher positioning accuracy than traditional fingerprint-matching positioning algorithms. Then, aiming at the problem of large fluctuations and poor continuity of fingerprint positioning results, a particle filter algorithm with an adaptive update of state parameters is proposed. The algorithm effectively integrates microelectromechanical systems (MEMS) sensor information in the smart phone and the structured spatial environment information, improves the freedom and positioning accuracy of pedestrian positioning, and achieves sub-meter-level stable absolute pedestrian positioning. Finally, in a test environment of about 800 m2, through a large number of experiments, compared with the millimeter-level precision optical dynamic calibration system, 94.2% of the positioning error is better than 1 m, and the average positioning error is 0.41 m. The results show that the system can provide high-precision and high-reliability location services and has great application and promotion value.

Modeling Vehicle–Pedestrian Interactions using a Nonprobabilistic Regression Approach

Understanding how vehicle drivers and pedestrians interact is key to identifying countermeasures that improve the safety of the interactions. As a result, techniques that can be used to evaluate the effectiveness of safety countermeasures and traffic control devices without the need to wait for the availability of crash data are needed. Using video, the interactions between right-turning vehicles and conflicting pedestrians were documented and quantified using vehicle and pedestrian position timestamps. Interactions documented were purposely narrow in scope to obtain a controlled dataset. Logged timestamps enabled the calculation of values such as time to complete a right turn and time for a pedestrian to reach a critical conflict point when a vehicle initiated a right turn. A nonprobabilistic regression model explaining the relationship between the calculated values was created. The model described the expected behavior of right-turning drivers: when drivers perceive the possibility of a pedestrian reaching a critical conflict point at the same time as them, they will modify their behavior, even if not coming to a complete stop. This behavior is not a surprise and has been previously documented in the literature. The primary contribution of this research is demonstrating that by analyzing a narrow set of interactions, clear and simple models that mostly explain the interactions between right-turning vehicles and pedestrians can be developed using nonprobabilistic linear regression techniques. An argument is made that the model parameters can be used to evaluate the effectiveness of traffic control devices.

Vision‐based approach for predicting the probability of vehicle–pedestrian collisions at intersections

Road accidents impose serious problems on society. Possible collisions between vehicles and pedestrians must be detected before they occur so that a timely warning may be issued. By using the vision-based approach, this study presents an effective and efficient algorithm to estimate the vehicle-pedestrian collision probability at intersections. The real-time trajectories and movement parameters (position, speed, acceleration or direction) of vehicles and pedestrians are obtained based on state-of-the-art detection and tracking algorithm which include background subtraction method, faster regions with convolutional neural networks and optical flow method. To find the appropriate time to identify the latent collision risk for calculating the collision probability, this study defines the critical time based on different collision patterns of perception-reaction failure and evasive action failure. In addition, based on discrete acceleration and discrete angle, the authors get different extended trajectories which can include most situation when the conflict happened. Trajectories generation probability are given by the discrete choice probability model based on the Logit model to get the accurate collision probability. Real-world video data is implemented to demonstrate the approach. This proposed collision prediction method can provide some important results for designing the intelligent pedestrian signal timing schemes at intersections.

Macroscopic Modelling of Pedestrian Flows Based on Conservation Law

Overcrowded of sidewalks can reduce the level of satisfaction of pedestrians. The large crowd on the sidewalks results in slowing down time of the pedestrian to arrive in their destination. This studt, focuses on pedestrian flows modelling using the macroscopic model. Numerical approximation of the macroscopic model is formulated as scalar hyperbolic conservation laws. The Lax-Wendroff scheme is used to discretize the equation of conservation laws. The simulation results show that the numerical approximation in term of density confirms the exact solution. In this simulation, the velocity function is obtained by curve fitting of observation data using linear regression method. The observation data, which are consist of velocity-density relation, are obtained from observation of pedestrian flows. The study case of this research conducts on the sidewalks of Braga Street, Bandung, Indonesia. There are two velocity functions used in the simulation, i.e. c 1 = 0.58206 + −0.94476ρ and c 2 = 0.59926 + −0.78052ρ, respectively. In performing the velocity function c 2, the pedestrian leader position is approximately 1 meter in front of the pedestrian leader using the velocity function c 1 at final time T = 20 seconds and T = 30 seconds. Overall, the numerical experiment shows that the pedestrian leader using the velocity functions c 2 is faster than by using c 1.

AONet: Active Offset Network for crowd flow prediction

Predicting crowd flow is of great importance to public safety and traffic management. The crowd flow is difficult to predict accurately and timely due to the uncertainty of the future positions. In this paper, we propose a novel Active Offset Network (AONet), in which ActiveGRU (Active Gate Recurrent Unit) is designed to predict the variation of pedestrians’ positions in the crowd flow. Its inner location-variant recurrent structure is implemented by utilizing convolution operation on low dimensional spatio-temporal sequences to obtain fractional offset locations. Afterwards, the sampling locations are determined by bilinear interpolation on fractional offset locations. Moreover, a probabilistic sparse strategy is introduced to reduce the links between sampling locations during supervised training. Finally, the experiments over popular benchmarks demonstrate that our method can actively characterize the future positions of pedestrians. Meanwhile, the performance of the proposed AONet is superior over state-of-art baselines with regard to both accuracy and computational savings.

A Testbed for GNSS-Based Positioning and Navigation Technologies in Smart Cities: The HANSEL Project

In urban contexts, the increasing density of electronic devices equipped with Global Navigation Satellite System (GNSS) receivers and complementary positioning technologies is attracting research and development efforts devoted to an improvement of the quality of life towards the smart city paradigm. Vehicular and pedestrian positioning and navigation capabilities are among the major drivers for innovation in this process. Ultra-low-cost electronics such as smartphones and Internet of Things (IoT) sensors aim at providing accurate and reliable positioning solutions through a set of promising solutions. Among these, snapshot positioning allows to remotely perform the post-processing of GNSS signals in IoT sensor networks while Wi-Fi™ ranging and cooperative positioning provide auxiliary anchors of opportunity to enhance indoor/outdoor positioning capabilities. This paper presents an innovative platform to perform a centralised testing and assessment of such positioning and navigation technologies along with a set of results obtained in the context of the European project HANSEL, by relying on current network technologies and infrastructures (i.e., Wi-Fi™ and cellular connectivity).

Adaptive 3D Position Estimation of Pedestrians by Wearing One Ankle Sensor

Without the assistance of Global Positioning System (GPS), estimating the 3D position of indoor and outdoor pedestrians using a single wearable device is a daunting if not impossible task, especially if micro-sensors are used. Here, we present a novel 3D position estimation method using one integrated ankle sensing device, which consists of an accelerometer, a gyroscope, a magnetometer, and a barometer. Pedestrians' vertical position is estimated by fusing the acceleration and angular velocity as well as the height derived from barometer data. To estimate pedestrians' horizontal position, we proposed an adaptive multimodal stride length model and a multi-sensor-fusion-based heading angle estimation method, using the Pedestrian Dead Reckoning (PDR) mechanism. By introducing a vertical variable into the stride length model, this new model greatly improved the applicability and accuracy of PDR in horizontal position estimation. Based on this new model, PDR can be used not only for estimating the horizontal position of the pedestrian walking on flat areas but also for those walking up and down stairs. The effectiveness and applicability of our method in 3D position estimation have been demonstrated in several different experiments of indoor and outdoor scenes. The first is when a pedestrian walks on an indoor flat ground, following a spiral trajectory. The estimation accuracy of pedestrians' height position is 1.5cm and the ratio of horizontal walking estimation error is 1.02m with a total walking distance of 53.1m. The second is when a pedestrian walks up staircases from Floor 4 to Floor 8 in a building. The cumulative error of the estimated height is 0.23m with a total height of 14.4m, and the root mean square error of the estimated horizontal 2D position is 15cm compared to a total horizontal walking distance of 38.4m. The third case involves a pedestrian walking on audience stands of an outdoor stadium and returns to the starting position for a total distance of 94.4m. The estimation error from the sensing device is 0.92m in this case, with mean error of the estimated height position of each step at 3.3cm. Having the capability to provide centimeter-level position estimation of pedestrians, this sensing device can be applied for 3D body tracking and indoor/outdoor pedestrian positioning and navigation.

Pedestrian Navigation System with Trinal-IMUs for Drastic Motions.

The combination of biomechanics and inertial pedestrian navigation research provides a very promising approach for pedestrian positioning in environments where Global Positioning System (GPS) signal is unavailable. However, in practical applications such as fire rescue and indoor security, the inertial sensor-based pedestrian navigation system is facing various challenges, especially the step length estimation errors and heading drift in running and sprint. In this paper, a trinal-node, including two thigh-worn inertial measurement units (IMU) and one waist-worn IMU, based simultaneous localization and occupation grid mapping method is proposed. Specifically, the gait detection and segmentation are realized by the zero-crossing detection of the difference of thighs pitch angle. A piecewise function between the step length and the probability distribution of waist horizontal acceleration is established to achieve accurate step length estimation both in regular walking and drastic motions. In addition, the simultaneous localization and mapping method based on occupancy grids, which involves the historic trajectory to improve the pedestrian’s pose estimation is introduced. The experiments show that the proposed trinal-node pedestrian inertial odometer can identify and segment each gait cycle in the walking, running, and sprint. The average step length estimation error is no more than 3.58% of the total travel distance in the motion speed from 1.23 m/s to 3.92 m/s. In combination with the proposed simultaneous localization and mapping method based on the occupancy grid, the localization error is less than 5 m in a single-story building of 2643.2 m.

Autonomous Vehicle Source Enumeration Exploiting Non-Cooperative UAV in Software Defined Internet of Vehicles

The traffic congestion and accidents can be relieved by deploying the software defined internet of vehicles (SDN-IoV). However, the traffic of pedestrians and vehicles is particularly heavy near commercial streets and campuses. In particular scenarios, the SDN-IoV may not ensure the quality of service (QoS) for pedestrians and vehicles. In this paper, we construct a novel system architecture consisting of multiple non-cooperative unmanned aerial vehicles (UAVs) and a SDN-IoV. The non-cooperative UAV is equipped with an antenna array to receive the signals from the vehicles and pedestrians of SDN-IoV. In order to locate the positions of vehicles and pedestrians, two source enumeration methods are proposed in a complex SDN-IoV environment with color noise. The projection matrix of the low dimensional signal subspace is constructed by the proposed criterion based on signal subspace projection (SSP). The sequence of the projected difference values of the local covariance matrix is applied to estimate the number of vehicles and pedestrians. The eigenvalues can be grouped to construct different subspaces by the proposed eigen-subspace projection (ESP). By projecting a new covariance matrix into the eigen-subspaces, the variance of values represents the projection difference can be exploited to estimate the number of vehicles and pedestrians. Simulation results and real system test verify the validity of the two proposed methods by comparing them with the state-of-the-art methods. Both of the methods have excellent estimation performance especially in color noise.

Improved Position Accuracy of Foot-Mounted Inertial Sensor by Discrete Corrections from Vision-Based Fiducial Marker Tracking.

In this paper, we present a novel pedestrian indoor positioning system that uses sensor fusion between a foot-mounted inertial measurement unit (IMU) and a vision-based fiducial marker tracking system. The goal is to provide an after-action review for first responders during training exercises. The main contribution of this work comes from the observation that different walking types (e.g., forward walking, sideways walking, backward walking) lead to different levels of position and heading error. Our approach takes this into account when accumulating the error, thereby leading to more-accurate estimations. Through experimentation, we show the variation in error accumulation and the improvement in accuracy alter when and how often to activate the camera tracking system, leading to better balance between accuracy and power consumption overall. The IMU and vision-based systems are loosely coupled using an extended Kalman filter (EKF) to ensure accurate and unobstructed positioning computation. The motion model of the EKF is derived from the foot-mounted IMU data and the measurement model from the vision system. Existing indoor positioning systems for training exercises require extensive active infrastructure installation, which is not viable for exercises taking place in a remote area. With the use of passive infrastructure (i.e., fiducial markers), the positioning system can accurately track user position over a longer duration of time and can be easily integrated into the environment. We evaluated our system on an indoor trajectory of 250 m. Results show that even with discrete corrections, near a meter level of accuracy can be achieved. Our proposed system attains the positioning accuracy of 0.55 m for a forward walk, 1.05 m for a backward walk, and 1.68 m for a sideways walk with a 90% confidence level.

Real-Time Height Measurement for Moving Pedestrians

Height measurement for moving pedestrians is quite significant in many scenarios, such as pedestrian positioning, criminal suspect tracking, and virtual reality. Although some existing height measurement methods can detect the height of the static people, it is hard to measure height accurately for moving pedestrians. Considering the height fluctuations in dynamic situation, this paper proposes a real-time height measurement based on the Time-of-Flight (TOF) camera. Depth images in a continuous sequence are addressed to obtain the real-time height of the pedestrian with moving. Firstly, a normalization equation is presented to convert the depth image into the grey image for a lower time cost and better performance. Secondly, a difference-particle swarm optimization (D-PSO) algorithm is proposed to remove the complex background and reduce the noises. Thirdly, a segmentation algorithm based on the maximally stable extremal regions (MSERs) is introduced to extract the pedestrian head region. Then, a novel multilayer iterative average algorithm (MLIA) is developed for obtaining the height of dynamic pedestrians. Finally, Kalman filtering is used to improve the measurement accuracy by combining the current measurement and the height at the last moment. In addition, the VICON system is adopted as the ground truth to verify the proposed method, and the result shows that our method can accurately measure the real-time height of moving pedestrians.

Stereo Vision by Combination of Machine-Learning Techniques for Pedestrian Detection at Intersections Utilizing Surround-View Cameras

The frequency of pedestrian traffic accidents continues to increase in Japan. Thus, a driver assistance system is expected to reduce the number of accidents. However, it is difficult for the current environmental recognition sensors to detect crossing pedestrians when turning at intersections, owing to the field of view and the cost. We propose a pedestrian detection system that utilizes surround-view fisheye cameras with a wide field of view. The system can be realized at low cost if the fisheye cameras are already equipped. It is necessary to detect the positioning of pedestrians accurately because more precise prediction of future collision points is required at intersections. Stereo vision is suitable for this purpose. However, there are some concerns regarding realizing stereo vision using fisheye cameras due to the distortion of the lens, asynchronous capturing, and fluctuating camera postures. As a countermeasure, we propose a novel method combining various machine-learning techniques. The D-Brief with histogram of oriented gradients and normalized cross-correlation are combined by a support-vector machine for stereo matching. A random forest was adopted to discriminate the pedestrians from noise in the 3D reconstructed point cloud. We evaluated this for images of crossing pedestrians at actual intersections. A tracking rate of 96.0% was achieved as the evaluation result. It was verified that this algorithm can accurately detect a pedestrian with an average position error of 0.17 m.

A High-Precision and Low-Cost IMU-Based Indoor Pedestrian Positioning Technique

Indoor pedestrian positioning has been widely used in applications such as fire rescue and indoor navigation. Compared with other indoor positioning technologies such as Wi-Fi and UWB, inertial measurement unit (IMU) based indoor positioning does not require external facilities and has lower cost. However, the major issue of IMU-based indoor positioning is that the inertial sensors exhibit errors and the errors accumulates, which affects the positioning precision. Existing IMU-based indoor positioning techniques have reduced the accumulated error, but still face several issues in positioning precision and system cost. In this paper, we propose a new IMU-based indoor pedestrian positioning technique. It adopts motion speed based adaptive error compensation and step detection based up/downstairs tracking to improve the positioning precision without using additional sensors. Compared with the existing techniques, the proposed technique achieves higher positioning precision (down to 0.11% for the closed error, 0.15% for the distance error and 0.52% for the height error) with less number of sensors (only accelerator and gyroscope) to lower the cost.

Multi-Stage Pedestrian Positioning Using Filtered WiFi Scanner Data in an Urban Road Environment.

Since widespread applications of wireless sensors networks, low-speed traffic positioning based on the received signal strength indicator (RSSI) from personal devices with WiFi broadcasts has attracted considerable attention. This study presents a new range-based localization method for outdoor pedestrian positioning by using the combination of offline RSSI distance estimation and real-time continuous position fitting, which can achieve high-position accuracy in the urban road environment. At the offline stage, the piecewise polynomial regression model (PPRM) is proposed to formulate the Euclidean distance between the targets and WiFi scanners by replacing the common propagation model (PM). The online stage includes three procedures. Firstly, a constant velocity Kalman filter (CVKF) is developed to smooth the real-time RSSI time series and estimate the target-detector distance. Then, a least squares Taylor series expansion (LS-TSE) is developed to calculate the actual 2-dimensional coordinate with the replacement of existing trilateral localization. Thirdly, a trajectory-based technique of the unscented Kalman filter (UKF) is introduced to smooth estimated positioning points. In tests that used field scenarios from Guangzhou, China, the experiments demonstrate that the combined CVKF and PPRM can achieve the highly accurate distance estimator of <1.98 m error with the probability of 90% or larger, which outperforms the existing propagation model. In addition, the online method can achieve average positioning error of 1.67 m with the much better than classical methods.