A Novel Intelligent Approach to Lane-Change Behavior Prediction for Intelligent and Connected Vehicles

The prediction of lane-change behavior is a challenging issue in intelligent and connected vehicles (ICVs), which can help vehicles predict in advance and change lanes safely. In this paper, a novel intelligent approach, which considering both the driving style-based lane-change environment and the driving trajectory-related parameters of the ICV and surrounding vehicles, is proposed to predict the lane-change behaviors for ICVs. By analyzing the characteristics of the lane-change behavior of the vehicle, a modified dataset for the prediction of lane-change behavior was established based on the Next-Generation Simulation (NGSIM) dataset, which is made up of real vehicle trajectories collected by camera. In the proposed approach, the hidden Markov model (HMM)-based model is designed to judge whether the current environment is suitable for lane change according to the driving environment parameters around the vehicle; then according to the driving state of the vehicle, a learning-based prediction-then-judgment model is proposed and designed to realize the prediction of the ICV's lane-change behavior. Experiments are implemented by using the modified dataset. From the experimental results, the lane-change probability value on the target lane in the truth of the lane-change behavior calculated by the designed HMM-based model is basically above 0.5, indicating that the model can make a more accurate judgment on the surrounding lane-change environment. The proposed learning-based prediction-then-judgment model has an accuracy of 99.32% for the prediction of lane-change behavior, and the accuracy of the lane-change detection algorithm in the model is 99.56%. The experimental results show that the proposed approach has a good performance in the prediction of lane-change behavior, which could effectively assist ICVs to change lanes safely.


Introduction
e intelligent and connected vehicle (ICV) [1] integrates modern communication and network technology and has environment perception, intelligent decision-making, and collaborative control functions. It can achieve safe, efficient, comfortable, and energy-saving driving and realize a new generation of vehicle that replaces humans [2].
Lane-change behavior detection and prediction plays an important role in the ICV technology. During the driving of the vehicle, the current driving environment may be misjudged due to the occlusion of surrounding vehicles or the driver's inattention, resulting in greater safety hazards. erefore, the sensor and communication technology can assist the ICV to perceive and judge the surrounding environment and the state of the vehicle and combined with artificial intelligence technology can predict the lanechanging behavior, thereby improving the driving safety.
Many methods have been proposed for lane-change detection and prediction, in which the main technical means and data sources used can be summarized as trajectories, steering wheel, surrounding environment, driving style, computer vision and roadside LiDAR, etc., as shown in Table 1. Lane-change behavior detection methods based on trajectory data are proposed in Refs. [3][4][5][6][7][8], such as fuzzy logic [3], support vector machine (SVM) [4], long short-term memory network and convolutional neural network (LSTM-CNN) [5], maneuver classification [6], and hidden Markov model [7,8]. Panichpapiboon and Leakkaw explore an approach to detect lane-change behavior using steering wheel angles extracted from the smart phone [9]. Zheng and Hansen propose an approach to detect lane-change behavior using the steering angle signal from CAN-bus [10]. Ali et al. propose a wavelet transform (WT)-based method to detect failed lane-changing attempts and used the random parameter binary logic model to study how the connected environment affects related parameters [11]. Woo et al. present a method to determine the driving styles and use the result to detect the lane-change behavior [12]. Nguyen et al. introduce a vision-based lane and vehicle detection approach for the lane-change assistant system [13]. Wang et al. present a method to detect lane-change behavior based on candidate lane markings [14]. Wei et al. develop a computer vision system to detect the lane-change behavior [15]. Cui et al. develop the methods to detect and predict lane-change behavior using vehicle trajectories from roadside LiDAR data [17]. Xu et al. present a V2X-based lane-change prediction model using vehicle trajectories [18]. Zhang and Fu present a lane-change intention detection method using motion parameters of the vehicle and surrounding vehicles [19]. Gao et al. introduce a lane-change behavior detection approach using multiple differing modality data [20]. Jin et al. present an optimal lane-change timing prediction model based on the driver's habits [21]. Huang et al. present a trajectory planning and control approach based on user preferences [22]. Xing et al. propose a driving pattern analysis and motion prediction system that determines the trajectory according to user's preference [23]. Xing et al. develop a driver intention inference system for highway lane-change maneuvers [16]. Xing et al. present a leading vehicle trajectory prediction approach that considers different driving styles [24].
In the above studies, different methods and models using different technical means and considering the influence of different characteristic parameters have been designed and proposed, fully verified, and achieved great results. However, few studies have simultaneously considered the effects of the vehicle, environment and driver, and the relationship between them when the ICV changes lanes. In this paper, a novel intelligent approach combines the driving state of the vehicle, the surrounding driving environment, and the driving style is proposed to predict the lane-change behaviors for ICVs. First, based on the learning of the driving habits of manual drivers, the current lane-change environment is judged according to the driving state of surrounding vehicles. If the current environment is suitable for lane change, then the vehicle driving state parameters are predicted, and the lane-change behavior detection method is proposed to judge the predicted value, so as to predict the lane-change behavior.
e main contributions can be summarized as follows.
(i) According to the relevant characteristic parameters of the vehicle lane change, the NGSIM dataset is processed and analyzed, so that a modified dataset for the lane-change behavior prediction is established (ii) Based on the driving habits of manual drivers, a HMM-based model is designed to judge whether the current surrounding environment of the vehicle is suitable for the lane change (iii) Based on the analysis of lane-change behavior characteristics, a prediction model based on LSTM and lane-change feature judgment method is proposed to predict the state parameters of the vehicle and determine whether it will change lanes (iv) A novel approach based on intelligent and connected technology, which in combination with the driving style-based lane-change environment and the driving trajectory-related parameters of the vehicle and surrounding vehicles, is proposed and performed on the established dataset to predict the lane-change behavior of vehicles e rest of the paper is organized as follows. In Section 2, the establishment process of the dataset is described. In Section 3, on the basis of fully analyzing the characteristics of lane-change behavior, the proposed approach to lanechange behavior prediction is introduced in detail. Section 4 gives the experimental results and analysis of the proposed approach performed on the modified dataset. Section 5 concludes the research work and presents the future work.

Dataset Establishment.
In this paper, the NGSIM dataset is processed to obtain the vehicle's trajectory and surrounding driving environment data, so as to combine the  [25]. In the NGSIM, I-80 and US-101 are the datasets collected in highway, which are studied in this paper. As shown in Figure 1, both I-80 and US-101 consist of five main lanes, one distribution lane, one on-ramp, and one off-ramp (the off-ramp of I-80 is not located within the study area). In I-80, the 1650-foot-long study area is divided into seven sub-areas by seven cameras to record the relevant data, while in US-101, the 2100-foot-long study area is divided into eight sub-areas by eight cameras. e dataset contains the trajectory data of all vehicles in the study area during the recorded time period.

Data Preprocessing.
In order to analyze the characteristics of the lane-change behavior, characteristic parameters such as the coordinates and velocity of the vehicles are extracted from the NGSIM dataset. e coordinates of the ramp and the most adjacent lane have a large overlap, which will cause great interference to the study. erefore, the data related to the ramp and the most adjacent lane are eliminated in the study. To further analyze the influence of the surrounding lane-change environment on lane-change behavior and the relationship between them, the distance between the vehicle and the front and rear vehicles on the current lane and adjacent lanes is calculated. e lateral speed of the vehicle is also calculated in order to predict the lane-change behavior.
e complete data of 92 lanechanging vehicles, a total of 92932 frames, are finally screened out and processed; then a modified dataset is established.
In the processed data, the surrounding lane-change environment at the time of a certain vehicle lane-change frame is shown in Figure 2. e range of lane coordinates calculated according to the data in the processed dataset is shown in Table 2.

Analysis of Lane-Change Characteristics.
According to the study of Balal et al. [26], the main characteristics that affect drivers' lane change are D ft , D pft , D pt , D pc , and V c . Lane-change behaviors can be divided into left-lane change and right-lane change, so D ft , D pft , and D pt can be divided into D fl , D fr , D pfl , D pfr , D pl , and D pr , which were defined in Table 3. e typical lane-change scenario taking the right-lane change as an example can be described in Figure 3.

Intelligent Prediction Approach.
Based on the analysis of lane-change characteristics in real scene dataset, an intelligent prediction approach is proposed and established, in which the HMM-based model is used to judge the lane-change conditions, LSTM-based model is used to predict the current vehicle motion data that are suitable for change lanes, and then the designed lane-change detection algorithm is performed to complete the lane-change behavior prediction.

HMM-Based Lane-Change Environment Judgment.
HMM can be used to predict the probability of whether a vehicle changes lanes [27][28][29].
e vehicle and surrounding vehicles' driving state determines to a large extent whether the vehicle has the conditions for changing lanes. In this paper, based on the analysis of vehicle lane-change characteristics, eight parameters, D fl , D fr , D pfl , D pfr , D pl , D pr , D pc , and V c , are selected as observations to judge the surrounding lane-change environment.
As shown in Figure   As shown in Figure 4, the parameters of designed model can be defined as follows: where A means the state transition probability matrix, in which a ij is the probability of transition to state h j at time B represents the observation probability matrix, in which b j(M) is the probability of generating the observation O M under the condition that time T is in state h j : π indicates the initial state probability distribution, in which π i is the probability of being in state h i at time t � 1: Computational Intelligence and Neuroscience π � π i , In this paper, the dataset contains the observation sequence and the corresponding state sequence. erefore, the supervised learning method can be used to estimate parameters of HMM. e maximum likelihood estimation method is used, and the specific method is as follows: (1) Estimate the transition probability. Assume that the frequency of the sample at time t in state i and transition to state j at time t + 1 is A ij , then the estimation of state transition probability a ij is as follows:    Computational Intelligence and Neuroscience (2) Estimate the probability of observation. Assume that the frequency of the sample state is j and the observation is M is B jM , then the estimation of the probability b j(M) that the state is j and the observation is M is as follows: (3) Estimate the initial state probability. e estimate π i of the initial state probability π i is the frequency at which the initial state is h i in the sample.
After the model parameters are determined, using the forward probability and the backward probability, given the model λ and the observation O, the probability of being in the state h i at time t can be obtained: From the definition of forward probability α t (i) and backward probability β t (i), en

LSTM-Based Vehicle Trajectory Prediction.
After judging the surrounding lane-change environment, the data suitable for lane change would be screened out to predict the lane-change behavior. An LSTM [30] model is designed to predict the current vehicle motion data. e structure of an LSTM block [31] is shown as Figure 5, in which f i is the input activation function, f o is the output activation function, and f g is the gate activation function. At time t, x t is the input, h t is the hidden layer state, it is the output state of input gate, f t is the output state of forget gate, and o t is the output state of output gate, which can be expressed as follows: where w x i , w x f , and w x o are the input weight matrices; w h i , w h f , and w h o who are the feedback weight matrices; and b i , b f , and b o are the bias vectors.  Observation states Computational Intelligence and Neuroscience e intermediate states at time t are as follows: the output state C in t corresponding to the input function, the output state C t corresponding to the output function, and the output state h t corresponding to the hidden layer.
where w x c , w h c , and b C in are the input weight matrix and the corresponding bias vector, respectively. C in t , as the input function, the output state at time t will participate in the overall update of the input state at time t together with the output state i t of the input gate at time t. As the output state of the input function at time t, C in t participates in the overall update of the input state at time t together with the output state i t of the input gate at time t. At time t, through the new input and state feedback at previous time, the entire LSTM unit is updated, including the update of C t and h t : In the update process of each gate function and the output state of the entire unit, the key information in the input feature is retained and transferred through the forget gate function and the transfer of the state.

e Lane-Change Behavior Prediction
Approach. e predicted data with conditions for lane change are used to determine whether the current vehicle will change lanes through the lane-change detection algorithm, which can be described as shown in Algorithm 1. An optimal sampling interval length (δ sampling period) is obtained according to the data training, and then the velocity data on x-axis are used as input according to the obtained sampling interval length. In the process of lane-change detection, first,

Evaluation Metrics.
When evaluating the prediction results, the confusion matrix definition of the prediction results is shown in Table 4. Accuracy, precision, recall, and F1 value are usually used as evaluation indicators [34] for learning-based classification and prediction models. Among them, the accuracy represents the proportion of the sample size correctly classified in the total sample size, which can be defined as follows: Precision (P), which indicates the proportion of samples with the correct class label among the samples of a particular class found by the classifier, can be defined as follows: e recall (R) represents the classifier's ability to find samples of a certain category, which can be defined as follows: e F1 value is a comprehensive index that considers the balance between precision and recall, which can be defined as follows: e closer the F1 value is to 1, the better the effect.

Performance of LSTM-Based Model.
In order to predict the lane-change behavior of the ICV, the LSTMbased model is designed to predict the lane-change-related motion data (lateral velocity) at the next moment. e dataset is divided into training set and test set at a ratio of 2 to 1 to verify the performance of the designed model. e loss curve of the training process is shown in Figure 8, in which the loss value is stable at around 2.32E-05. e prediction result of designed LSTM is shown in Figure 9, in which the blue line represents original data and yellow line and green line indicate the prediction result of the training set and the test set, respectively. e root mean square error (RMSE) of the prediction is 0.37 m/s for the training set and 0.68 m/s for the test set. From the prediction results in the figure, it can be found that the overall prediction error of the designed model is small, and the prediction error is greater when the data have large and sudden changes than when the data are flat. e maximum prediction error of the dataset is 5.5668 m/s (the original data is 50.0055 m/s).

Performance of the Detection Algorithm.
e designed lane-change behavior detection algorithm was performed on the established dataset to verify its detection effect on lanechange behavior. e dataset is divided into training set and test set at a ratio of 2 to 1, the experimental result of detection is shown in Table 5, and the confusion matrix of it is shown in Figure 10.
It can be seen from the experimental result of detection that the P of left-lane-change detection has reached 100% and R of it is 83.05%, the P of right-lane-change detection is 90.91% and R of it is 100%, while P and R of nonlane change Taking entire driving process of vehicle 2458 as an example, the detection result is shown in Figure 11. It can be seen from the figure that vehicle 2458 made a lane change at t � 600 during the whole process, and its lateral velocity has an obvious acceleration process. e designed detection algorithm accurately detects the lane-change behavior and calculates the lane-change process that is between t 1 � 591 and t 2 � 641.
Input: e velocity data on x-axis in δ sampling periods, V x ; Output: e predicted lane-change behavior, behavior; e beginning and ending time of the predicted lane-change behavior, t 1 , (6) for all zero crossing value m � 1 to length do (9) for all training object (V m , behavior) ∈ I and test object z � (V m , 2 between z and every object (Vm, behavior) ∈ I; (11) Select I z ⊆ I, the set of k closest training objects to z; return behavior   Figure 6: e structure of lane-change behavior prediction approach.

Performance of the Lane-Change Behavior Prediction
Approach. Finally, the proposed prediction approach is performed on the established dataset to verify the effect of the approach. e proposed lane-change behavior detection algorithm is performed on the filtered prediction data that meets the lane-change conditions. e dataset is also divided into training set and test set at a ratio of 2 to 1, the experimental result of prediction is shown in Table 6, and the confusion matrix of prediction result is shown in Figure 12.    could effectively provide vehicle lane-change information to assist the ICV in safe lane change.
Taking the entire driving process of vehicle 2458 as an example, the prediction result is shown in Figure 13. It can be seen that the predicted lateral velocity and the truth of lateral velocity are basically consistent in value and trend. e designed prediction approach accurately predicts the lane-change behavior and calculates the lane-change process that is between t 1 � 590 and t 2 � 776. e predicted time interval of the lane-change process is longer than that calculated by the detection algorithm, which is because that when the lateral velocity value fluctuates around 0, the predicted lateral velocity value fluctuates slightly and is less than 0. e prediction approach can still accurately predict the lane-change behavior and the time interval of lane change.   On the basis of the fusion of the above models and algorithm, the proposed intelligent prediction approach is completed, result shows that the accuracy of the prediction approach on the established dataset is 99.32%, and the time interval of the vehicle lane change can be calculated accurately. e experimental result indicates that the proposed prediction approach could effectively provide vehicle lane-change information to assist the ICV in safe lane-change and has the potentials for application in actual intelligent and connected environment for ICVs.

Conclusions
Since the proposed approach is postprocessing with measured data, its real-time performance in practical Computational Intelligence and Neuroscience applications needs to be further verified. In future work, the proposed approach can be deployed on mobile terminals for real-time testing, and its accuracy and real-time performance can be further improved.
Data Availability e modified NGSIM data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest
e authors declare that there are no conflicts of interest regarding the publication of this paper.