Trajectory Uncertainties Research Articles

Advancing Vehicle Trajectory Prediction: A Probabilistic Approach Using Combined Sequential Models

Abstract This paper addresses the critical need to quantify vehicle trajectory uncertainty in autonomous driving under environmental variability. We focus on predicting the posterior distribution of vehicle trajectories over a fixed horizon, given an initial state and a sequence of actions. We propose and compare three approaches: a probabilistic seq2seq model based on stochastic variational Gaussian processes, sequential Monte Carlo simulation with a single-step Gaussian process model, and a hybrid model that leverages the strengths of both methods. Each approach incorporates a baseline vehicle kinematics model to enhance stability and convergence. We evaluate these methods using a dataset generated from the CARLA simulator, assessing both point error metrics and probabilistic prediction metrics. This research introduces novel approaches to quantifying vehicle trajectory uncertainty through various uncertainty quantification techniques, with the goal of improving the safety and reliability of autonomous vehicle control systems.

Kinematic calibration of industrial robot using Bayesian modeling framework

Positioning accuracy of an end-effector is a crucial metric for evaluating industrial robot performance. Uncertainties in joint angles and joint backlash deviate actual angles from the designed nominal values to negate positioning accuracy. Most existing parameter identification methods overlook or not properly account for such uncertainties, leading to usually overconfident identification results. To this gap, the present study introduces a kinematic calibration methodology employing Bayesian parameter estimation to achieve identification of joint variables. New formulas based on data features of industrial robots for constructing the likelihood function are proposed, and model selection is applied to assess various likelihood functions for a tradeoff balance between complexity and accuracy. To evaluate the robustness of the proposed approach, identification of joint variables is conducted under different measurement noises. The position response of kinematic model is predicted based on the identified joint uncertainty information. The efficacy is verified through rigorous scrutiny involving both a numerical example and an engineering application. Results indicate that the proposed method exhibits satisfactory kinematic parameter identification accuracy and robustness. In addition, the uncertainty of parameters can be measured and the prediction of trajectory uncertainty intervals is realized simultaneously, which promotes the application of industrial robots in high-precision scenes.

Developing an evidence-and ethics-informed intervention for moral distress.

The global pandemic has intensified the risk of moral distress due to increased demands on already limited human resources and uncertainty of the pandemic's trajectory. Nurses commonly experience moral distress: a conflict between the morally correct action and what they are required or capable of doing. Effective moral distress interventions are rare. For this reason, our team conducted a multi-phase research study to develop a moral distress intervention for pediatric critical care nurses. In this article, we discuss our multi-phase approach to develop a moral distress intervention-proactive, interdisciplinary meeting. Our proposed intervention is a sequential compilation of empirical work couched within a relational ethics lens thus should point to enhanced potential for intervention effectiveness.

Predicting Air Traffic Congestion under Uncertain Adverse Weather

This paper presents an approach for integrating uncertainty information in air traffic flow management at the tactical phase. In particular, probabilistic methodologies to predict sector demand and sector congestion under adverse weather in a time horizon of 1.5 h are developed. Two sources of uncertainty are considered: the meteorological uncertainty inherent to the forecasting process and the uncertainty in the take-off time. An ensemble approach is adopted to characterize both uncertainty sources. The methodologies rely on a trajectory predictor able to generate an ensemble of 4D trajectories that provides a measure of the trajectory uncertainty, each trajectory avoiding the storm cells encountered along the way. The core of the approach is the statistical processing of the ensemble of trajectories to obtain probabilistic entry and occupancy counts of each sector and their congestion status when the counts are compared to weather-dependent capacity values. A new criterion to assess the risk of sector overload, which takes into account the uncertainty, is also defined. The results are presented for a historical situation over the Austrian airspace on a day with significant convection.

Hierarchical vector transformer vehicle trajectories prediction with diffusion convolutional neural networks

In dynamic and interactive autonomous driving scenarios, accurately predicting the future movements of vehicle agents is crucial. However, current methods often fail to capture trajectory uncertainty, leading to limitations in trajectory prediction performance. To address these limitations, this paper introduces the hierarchical vector transformer diffusion model, a novel trajectory prediction method that prioritizes both speed and accuracy. The proposed model decomposes the traffic scene modeling into local patches and global interactions, allowing for the acquisition of relevant environmental and global information. Moreover, a local diffusion encoder is employed to effectively capture the aleatoric uncertainty. The proposed model utilizes an adaptive graph structure to exploit the spatial and temporal relationships inherent in the trajectory data. By employing a graph diffusion process, the model effectively captures dynamic features from the historical trajectory information. Moreover, the model demonstrates adaptability by dynamically adjusting to diverse trajectory data and scenarios, thereby enabling the generation of predicted trajectories that are uncertainty aware. This approach contributes to more effective and efficient modeling of dynamic autonomous driving scenarios. Experimental results demonstrate the superior speed and accuracy of the proposed method compared to existing approaches for trajectory prediction. The proposed method significantly enhances prediction accuracy, achieving results of ADE 0.68 and FDE 1.02 on the Argoverse dataset. In comparison to the baseline model, there are notable improvements in ADE and FDE by 0.03 and 0.06, respectively. It is noteworthy that this method also reduces the inference time by 7% when compared to the currently fastest method.

Characterization of Gauss–Markov stochastic sequences for mission analysis

In real scenarios, the spacecraft deviates from the intended paths owing to uncertainties in dynamics, navigation, and command actuation. Accurately quantifying these uncertainties is crucial for assessing the observability, collision risks, and mission viability. This issue is further magnified for CubeSats because they have limited control authority and thus require accurate dispersion estimates to avoid rejecting viable trajectories or selecting unviable ones. Trajectory uncertainties arise from random variables (e.g., measurement errors and drag coefficients) and processes (e.g., solar radiation pressure and low-thrust acceleration). Although random variables generally present minimal computational complexity, handling stochastic processes can be challenging because of their noisy dynamics. Nonetheless, accurately modeling these processes is essential, as they significantly influence the uncertain propagation of space trajectories, and an inadequate representation can result in either underestimation or overestimation of the stochastic characteristics associated with a given trajectory. This study addresses the gap in characterizing process uncertainties, represented as Gauss–Markov processes in mission analysis, by presenting models, evaluating derived quantities, and providing results on the impact of spacecraft trajectories. This study emphasizes the importance of accurately modeling random processes to properly characterize stochastic spacecraft paths.

Decisional Conflict Regarding Disease-Modifying Treatment Choices Among Patients with Mid-Stage Relapsing-Remitting Multiple Sclerosis.

Shared decision-making is critical in multiple sclerosis (MS) due to the uncertainty of the disease trajectory over time and the large number of treatment options with differing efficacy, safety and administration characteristics. The aim of this study was to assess patients' decisional conflict regarding the choice of a disease-modifying therapy and its associated factors in patients with mid-stage relapsing-remitting multiple sclerosis (RRMS). A multicenter, non-interventional study was conducted. Adult patients with a diagnosis of RRMS (2017 revised McDonald criteria) and disease duration of 3 to 8 years were included. The level of uncertainty experienced by a patient when faced with making a treatment choice was assessed using the 4-item Decisional Conflict Scale. A battery of patient-reported and clinician-rated measures was administered to obtain information on symptom severity, illness perception, illness-related uncertainty, regret, MS knowledge, risk taking behavior, preferred role in the decision-making process, cognition, and self-management. Patients were recruited during routine follow-up visits and completed all questionnaires online using electronic tablets at the hospital. A multivariate logistic regression analysis was conducted. A total of 201 patients were studied. Mean age (Standard deviation) was 38.7 (8.4) years and 74.1% were female. Median disease duration (Interquartile range) was 6.0 (4.0-7.0) years. Median EDSS score was 1.0 (0-2.0). Sixty-seven (33.3%) patients reported a decisional conflict. These patients had lower MS knowledge and more illness uncertainty, anxiety, depressive symptoms, fatigue, subjective symptom severity, a threatening illness perception, and poorer quality of life than their counterparts. Lack of decisional conflict was associated with MS knowledge (Odds ratio [OR]=1.195, 95% CI 1.045, 1.383, p=0.013), self-management (OR=1.049, 95% CI 1.013, 1.093, p=0.018), and regret after a healthcare decision (OR=0.860, 95% CI 0.756, 0.973, p=0.018) in the multivariate analysis. Decisional conflict regarding the selection of a disease-modifying therapy was a common phenomenon in patients with mid-stage RRMS. Identifying factors associated with decisional conflict may be useful to implement preventive strategies that help patients better understand their condition and strengthen their self-management resources.

LSTM-MLP BASED UNCERTAINTY MODELLING APPROACH FOR COMPLEX HUMAN INDOOR TRAJECTORY

Abstract. Modelling the movement uncertainty of human indoor trajectory consist of an essential part in promoting the performance of smart city related applications. At this stage, the existing uncertainty modelling algorithms usually take the constant sampling error or measurement error into consideration and cannot adapt well to the changeable human motion modes and complex handheld modes of smartphones. To fill this gap, this paper applied the Long Short-Term Memory (LSTM) network for continuous prediction of uncertainty error of human indoor trajectory with complex motion modes and detected indoor landmark points. The human motion information including handheld modes, walking speed, and heading information in extracted and fused with detected landmark points for reconstruction of human indoor trajectory under large-scale areas using Gradient Descent (GD) algorithm. In addition, the hybrid LSTM and Multilayer Perceptron (MLP) network is adopted for uncertainty error prediction, by considering both sampling error and measurement error in a specific time period, and the reconstructed trajectory with human motion features are modelled as the input vector for model training with the ground-truth uncertainty error as reference. Comprehensive experiments on real-world collected dataset indicate that the proposed LSTM-assisted uncertainty modelling algorithm has robust outperformance in uncertainty error prediction and uncertainty region definition compared with traditional uncertainty modelling approaches.

Collaborative Uncertainty Benefits Multi-Agent Multi-Modal Trajectory Forecasting.

In multi-modal multi-agent trajectory forecasting, two major challenges have not been fully tackled: 1) how to measure the uncertainty brought by the interaction module that causes correlations among the predicted trajectories of multiple agents; 2) how to rank the multiple predictions and select the optimal predicted trajectory. In order to handle the aforementioned challenges, this work first proposes a novel concept, collaborative uncertainty (CU), which models the uncertainty resulting from interaction modules. Then we build a general CU-aware regression framework with an original permutation-equivariant uncertainty estimator to do both tasks of regression and uncertainty estimation. Furthermore, we apply the proposed framework to current SOTA multi-agent multi-modal forecasting systems as a plugin module, which enables the SOTA systems to 1) estimate the uncertainty in the multi-agent multi-modal trajectory forecasting task; 2) rank the multiple predictions and select the optimal one based on the estimated uncertainty. We conduct extensive experiments on a synthetic dataset and two public large-scale multi-agent trajectory forecasting benchmarks. Experiments show that: 1) on the synthetic dataset, the CU-aware regression framework allows the model to appropriately approximate the ground-truth Laplace distribution; 2) on the multi-agent trajectory forecasting benchmarks, the CU-aware regression framework steadily helps SOTA systems improve their performances. Especially, the proposed framework helps VectorNet improve by 262 cm regarding the Final Displacement Error of the chosen optimal prediction on the nuScenes dataset; 3) in multi-agent multi-modal trajectory forecasting, prediction uncertainty is proportional to future stochasticity; 4) the estimated CU values are highly related to the interactive information among agents. The proposed framework can guide the development of more reliable and safer forecasting systems in the future. Our code is available at https://github.com/MediaBrain-SJTU/Collaborative-Uncertainty.

Analysis of UTM Tracking Performance for Conformance Monitoring via Hybrid SITL Monte Carlo Methods

Conformance monitoring supports UTM safety by observing if unmanned aircraft (UA) are adhering to declared operational intent. As a supporting system, robust cooperative tracking is critical. Nevertheless, tracking systems for UAS traffic management (UTM) are in an early stage and under-standardized, and existing literature hardly addresses the problem. To bridge this gap, this study aims to probabilistically evaluate the impact of the change in tracking performances on the effectiveness of conformance monitoring. We propose a Monte Carlo simulation-based method. To ensure a realistic simulation environment, we use a hybrid software-in-the-loop (SITL) scheme. The major uncertainties contributing to the stochastic evaluation are measured separately and are integrated into the final Monte Carlo simulation. Latency tests were conducted to assess the performance of different communication technologies for cooperative tracking. Flight technical error generation via SITL simulations and navigational system error generation based on flight experiments were employed to model UA trajectory uncertainty. Based on these tests, further Monte Carlo simulations were used to study the overall impacts of various tracking key performance indicators in UTM conformance monitoring. Results suggest that the extrapolation of UA position enables quicker non-conformance detection, but introduces greater variability in detection delay, and exacerbates the incidence of nuisance alerts and missed detections, particularly when latencies are high and velocity errors are severe. Recommendations for UA position update rates of ≥1 Hz remain consistent with previous studies, as investments in increasing the update rate do not lead to corresponding improvements in conformance monitoring performance according to simulation results.

Predicting aircraft trajectory uncertainties for terminal airspace design evaluation

The terminal airspace that surrounds an airport is the area with high flight density and complex structure. Aircraft are asked to follow the standard arrival and departure routes in terminal airspace, yet the actual trajectories may deviate due to air traffic control (ATC) instructions, pilots' decisions, surveillance and flying performance variations, etc. Predicting aircraft trajectories considering such uncertainties plays a crucial role in evaluating a redesign of the standard routes. Traditional simulation approaches for generating aircraft trajectories in a terminal airspace are cumbersome to use as it requires a detailed setup for each new scenario, while most existing data-driven methods can only be used in an airspace with historical trajectories, not applicable to new structure designs or other terminal areas. To fill in gap, in this paper, we develop a new model based on Multilayer Perceptron Neural Network (MLPNN) to predict aircraft trajectories with uncertainties for terminal airspace design evaluations. A key feature of the proposed model is that it is trained on existing standard routes yet it can be applied to new standard routes to generate trajectories. The enabler of the model's transferability is a novel input-and-output construction method for feature representations of raw trajectory data based on domain knowledge, including trajectory reconstruction, feature engineering, and output designing. After the input-and-output construction, a supervised learning model based on MLPNN is built to predict the standard deviations from the extracted features using historical trajectory data of existing standard routes. Once the model is built, trajectories with uncertainty can be simulated, through applying Gaussian distribution and exponential moving average algorithms, even on newly designed standard routes, where no aircraft have flown yet. Subsequently, new terminal airspace designs could be evaluated for their safety, efficiency, and environmental implications based on the simulated trajectories. The proposed model was tested on real-world operational data. Results showed that the model can quantify the characteristics of aircraft trajectories that are transferable across standard routes, and generate trajectories for new standard routes. We also demonstrated the proposed model on evaluating deficiencies on fuel consumption of actual arrival trajectories compared with the designed arrival routes. The generated trajectories showed 23%–37% more fuel consumption on average than the standard arrival routes in the terminal airspace of Hong Kong International Airport, which was validated with actual flight data.

Learning vehicle trajectory uncertainty

A novel approach for vehicle tracking using a hybrid adaptive Kalman filter is proposed. The filter utilizes recurrent neural networks to learn the vehicle’s geometrical and kinematic features, which are then used in a supervised learning model to determine the actual process noise covariance in the Kalman framework. This approach addresses the limitations of traditional linear Kalman filters, which can suffer from degraded performance due to uncertainty in the vehicle kinematic trajectory modeling. Our method is evaluated and compared to other adaptive filters using the Oxford RobotCar dataset, and has shown to be effective in accurately determining the process noise covariance in real-time scenarios. Overall, this approach can be implemented in other estimation problems to improve performance.

Experience and perspectives of end-of-life care discussion and physician orders for life-sustaining treatment of Korea (POLST-K): a cross-sectional study

BackgroundThis study aimed to identify the healthcare providers’ experience and perspectives toward end-of-life care decisions focusing on end-of-life discussion and physician’s order of life-sustaining treatment documentation in Korea which are major parts of the Life-Sustaining Treatment Act.MethodsA cross-sectional survey was conducted using a questionnaire developed by the authors. A total of 474 subjects—94 attending physicians, 87 resident physicians, and 293 nurses—participated in the survey, and the data analysis was performed in terms of frequency, percentage, mean and standard deviation using the SPSS 24.0 program.ResultsStudy results showed that respondents were aware of terminal illness and physician’s order of life-sustaining treatment in Korea well enough except for some details. Physicians reported uncertainty in terminal state diagnosis and disease trajectory as the most challenging. Study participants regarded factors (related to relationships and communications) on the healthcare providers’ side as the major impediment to end-of-life discussion. Study respondents suggested that simplification of the process and more staff are required to facilitate end-of-life discussion and documentation.ConclusionBased on the study results, adequate education and training for better end-of-life discussion are required for future practice. Also, a simple and clear procedure for completing a physician’s order of life-sustaining treatment in Korea should be prepared and legal and ethical advice would be required. Since the enactment of the Life-Sustaining Treatment Act, several revisions already have been made including disease categories, thus continuous education to update and support clinicians is also called for.

Vertical autofocus for the phase screen in a turbulent ionosphere

The performance of spaceborne synthetic aperture radars (SARs) is affected by the Earth’s ionosphere. In particular, the ionospheric turbulence causes phase perturbations of the SAR signals, which may lead to image distortions. A convenient way to model those phase perturbations is by means of a phase screen. The latter is an infinitesimally thin layer positioned at a certain elevation above the Earth’s surface. The radar signal acquires an instant perturbation once its trajectory intersects the screen. The trajectory is a ray between the antenna and the target, and the magnitude of the perturbation is equal to the screen density at the intersection point. The density is a bivariate function of the coordinates along the screen. The coordinates of a specific intersection point are determined by the ray itself, as well as the screen elevation. Thus, the magnitude of the phase perturbation explicitly depends on the screen elevation. Accordingly, to compensate for the resulting image distortions one should be able to determine the elevation of the screen. In the paper, we develop an algorithm of vertical autofocus that derives this elevation from the received SAR data, given a pair of point scatterers in the target area. The proposed algorithm exploits a modification of the coherent interferometric imaging that was previously employed to reduce the effect of phase errors due to the trajectory uncertainty. In our analysis, we highlight the differences between this case and transionospheric propagation.

A novel multi-ship collision probability estimation method considering data-driven quantification of trajectory uncertainty

The collision risk prediction is crucial for the safety management of maritime transportation. Previous studies have primarily focused on the near-miss collision risk of two ships, yet the risk due to congestion caused by multiple ships is complex and also very challenging for safe management. In this paper, a novel multi-ship collision probability assessment method is proposed. To consider the uncertainty of ship motion, a deep learning multi-model integration method is proposed to predict the ship motion trajectory to quantify the time-varying stochasticity of ship motion. Then, a method for calculating collision probability is proposed based on Monte Carlo simulation by integrating importance sampling techniques with the cross-entropy method. The model is combined with the ship trajectory estimator to quantify the collision probability. To prove the validity of the developed model, a case study based on Victoria Harbour is illustrated. The results show that this method can provide early warning of multi-ship collision risk. The method can offer an important basis for maritime collision risk monitoring and ship navigation risk assessment for maritime administration departments and shore-based center of maritime autonomous ships.

Spatiotemporal Attention-Based Pedestrian Trajectory Prediction Considering Traffic-Actor Interaction

Accurate pedestrian trajectory prediction is significant and challenging for traffic-actor protection and intelligent driving. However, most of the existing methods focus on pedestrian-pedestrian interaction ignoring the pedestrian-vehicle interaction and containing much ineffective interaction. To tackle the problem, this paper proposes a spatiotemporal LSTM attention network (STLAN) for pedestrian trajectory prediction based on a spatiotemporal dual attention module (STDAM). First, a pedestrian social force model is established to analyze the effective pedestrian-pedestrian interaction and pedestrian-vehicle interaction in mixed traffic scene. Second, a pedestrian trajectory prediction network architecture is constructed to integrate temporal and spatial information. By temporal attention STDAM captures underlying dependencies between time steps in individual pedestrian, assigning different weights to the input sequence of it. By spatial attention STDAM extracts latent interaction between traffic actors, assigning different weights to every traffic actor. In addition, random noise is introduced to simulate the uncertainty of pedestrian trajectory. Experimental results on a real-world traffic actors dataset show that STLAN outperforms several mainstream methods. Specifically, STLAN achieves 0.17/0.11 ADE/FDE and 3.79/0.38 AMD/AMV in the representative pedestrian-pedestrian interaction scene, with 0.15/0.17 ADE/FDE and 2.36/0.21 AMD/AMV in pedestrian-vehicle interaction scene. Furthermore, ablation experiments validate the effectiveness of STDAM in improving trajectory prediction accuracy.

Clearance-induced contact trajectory uncertainty of angular contact ball bearing under coupling operating condition

Clearance-induced contact trajectory uncertainty of angular contact ball bearing under coupling operating condition

A Framework of Improving Human Demonstration Efficiency for Goal-Directed Robot Skill Learning

Robot learning from humans allows robots to automatically adjust to stochastic and dynamic environments by learning from nontechnical end user’s demonstrations, which is best known as robot programming by demonstration, robot learning from demonstration, apprenticeship learning, and imitation learning. Although most of those methods are probabilistic, and their performances intensively depend on the demonstrated data, measuring and evaluating human demonstrations are rarely investigated. A poorly demonstrated data set with useless prior knowledge or redundant demonstrations increases the complexity and time cost of robot learning. To solve these problems, a goal-directed robot skill learning framework named GPm-MOGP is presented. It 1) decides when and where to add a new demonstration by calculating the trajectory uncertainty; 2) determines which demonstration is useless or redundant by Kullback–Leibler (KL) divergence; 3) implements robot skill learning with a minimum number of demonstrations using a multioutput Gaussian process; and 4) learns orientation uncertainty and representation by combining logarithmic and exponential maps. The proposed framework significantly reduces the demonstrated effort of nontechnical end users who lack an understanding of how and what the robot learns during the demonstrating process. To evaluate the proposed framework, a pick-and-place experiment was designed with five unseen goals to verify the effectiveness of our methods. This experiment is well illustrated with two phases: 1) demonstration efficiency and 2) skill representation and reproduction. The results indicate an improvement of 60% in human demonstration efficiency, compared to common learning from demonstrations (LfD) applications that require at least ten demonstrations, and the robot average success rate of pick-and-place task reaches 85%.

A deep-learning approach for modelling pedestrian movement uncertainty in large- scale indoor areas

• Indoor trajectory reconstruction model using limited reference points is proposed. • Uncertainty prediction model uses the combination of 1D-CNN and LSTM is proposed. • Enhanced training dataset contains sampling and measurement errors is generated. • Uncertainty region considers both sampling and measurement errors is established. Modelling pedestrian movement uncertainty in complex urban environments is regarded as a meaningful and challenging task regarding the promotion of geospatial data mining and analysis. However, the traditional uncertainty prediction model only takes the movement distance or speed into consideration and is not able to adapt well to time-varying measurement errors. In this paper, a deep-learning framework is proposed for modelling pedestrian movement uncertainty in large-scale indoor areas, in which a hybrid deep-learning model combines a one-dimensional Convolutional Neural Network (1D-CNN) with a long short-term memory (LSTM) network is proposed for enhancing feature extraction performance and reducing time correlation errors. The proposed framework takes human motion related measurement features into consideration, in which the moving step-length and heading information during a time period are also reconstructed and modelled as the input to the deep-learning model. Compared with state-of-art algorithms applied to different real-world trajectory datasets, the proposed deep-learning approach demonstrates much better performance of uncertainty region prediction, including the different indexes (Euclidean error distance, completeness and density) This study has leaded to the provision of an effective and practical framework for modelling trajectory uncertainty of the pedestrian in challenging urban environments, and which is expected to benefit smart city and spatial perception related applications.

A V2P Collision Risk Warning Method based on LSTM in IOV

With the evolution of communication networks, the Internet of Vehicles (IOV) continues to accelerate the safe and rapid development of autonomous vehicles. Vehicle-to-Pedestrian (V2P) communication is a key technology in autonomous vehicles and a potential solution to realize collaborative intelligence between vehicles and pedestrians. However, the existing V2P communication early warning system does not consider the uncertainty of pedestrian trajectory, and the determination of the collision area is limited to a single point, resulting in an inaccurate system judgment and limited improvement of traffic efficiency. This paper designs a new autonomous-oriented V2P communication network architecture and completes a V2P communication early warning system based on Long Range (LoRa). A V2P anticollision model is established, and a new V2P collision risk early warning method is proposed. In this method, danger index is introduced into the early warning of collision between pedestrian and vehicle. The long short-term memory (LSTM) artificial neural network is used to predict the pedestrian’s trajectory, so as to deduce the pedestrian-vehicle collision risk area when the pedestrian trajectory is uncertain. Meanwhile, the confidence probability is used to judge whether the pedestrian and vehicle are warned. The simulation shows that the V2P collision risk warning method proposed in this paper has good performance, which can accurately warn the pedestrian and vehicle under different vehicle speeds and Global Positioning System (GPS) positioning errors. At the same time, it reflects the characteristics of intelligence brought by using LSTM methods. Using the V2P communication early warning system based on LoRa to verify the experimental results show that when the GPS positioning accuracy is submeter level, the prediction accuracy is greater than 98%. The results of the proposed method show good performance and high detection rate.