Abstract: A system and method for predicting the near-term trajectory of a moving obstacle sensed by an autonomous driving vehicle (ADV) is disclosed. The method applies neural networks such as a LSTM model to learn dynamic features of the moving obstacle's motion based on its past trajectory up to its current position and a CNN model to learn the semantic map features of the driving environment in a portion of an image map. From the learned dynamic features of the moving obstacle and the learned semantic map features of the environment, the method applies a neural network to iteratively predict the moving obstacle's positions for successive time points of a prediction interval. To predict the moving obstacle's position at the next time point from the currently predicted position, the methods may update the learned dynamic features of the moving obstacle based on its past trajectory up to the currently predicted position.

Abstract: A parking system for autonomous driving vehicles optimizes a solution to a parking problem. The ADV detects a parking lot and selects a parking space. The ADV defines constraints for the parking lot, parking space, and kinematic constraints of the ADV, and generates a plurality of potential parking paths to the parking space, taking into account the constraints of the parking lot, parking space, and kinematics of the ADV, but without taking into any obstacles that may be surrounding the ADV. The ADV determines a cost for traversing each of the parking paths. One or more least cost candidate paths are selected from the parking paths, then one or more candidate paths are eliminated based on obstacles surrounding the ADV. Remaining candidates can be analyzed using a quadratic optimization system. A best parking path can be selected from the remaining candidates to navigate the ADV to the parking space.

Abstract: Methods and systems for multimodal motion planning framework for autonomous driving vehicles are disclosed. In one embodiment, driving environment data of an autonomous vehicle is received, where the environment data includes a route segment. The route segment is segmented into a number of route sub-segments. A specific driving scenario is assigned to each of the route sub-segments, where each specific driving scenario is included in a set of driving scenarios. A first motion planning algorithm is assigned according to a first assigned driving scenario included in the set of driving scenarios. The first motion planning algorithm is invoked to generate a first set of trajectories. The autonomous vehicle is controlled based on the first set of trajectories.

Abstract: A system and method for predicting the near-term trajectory of a moving obstacle sensed by an autonomous driving vehicle (ADV) is disclosed. The method applies neural networks such as a LSTM model to learn dynamic features of the moving obstacle's motion based on its past trajectory up to its current position and a CNN model to learn the semantic map features of the driving environment in a portion of an image map. From the learned dynamic features of the moving obstacle and the learned semantic map features of the environment, the method applies a neural network to iteratively predict the moving obstacle's positions for successive time points of a prediction interval. To predict the moving obstacle's position at the next time point from the currently predicted position, the methods may update the learned dynamic features of the moving obstacle based on its past trajectory up to the currently predicted position.

Abstract: A method, apparatus, and system for generating an optimal path for an autonomous driving vehicle (ADV) are disclosed. The method includes receiving optimization inputs comprising an ADV starting state, a maximal lateral jerk, and static obstacle boundaries with respect to a reference line; receiving optimization constraints comprising constraints relating to the maximal lateral jerk and avoidance of one or more static obstacles; receiving a cost function associated with an optimization objective, the cost function comprising a first term relating to cumulative lateral distances, a second term relating to cumulative first order lateral rates of change, and a third term relating to cumulative second order lateral rates of change; generating planned ADV states as optimization results with nonlinear optimization, by minimizing a value of the cost function; and generating control signals to control the ADV based on the plurality of planned ADV states.

Abstract: A moving obstacle such as a vehicle within a proximity of an intersection and one or more exits of the intersection are identified. An obstacle state evolution of a spatial position of the moving obstacle over a period of time is determined. For each of the exits, an intersection exit encoding of the exit is determined based on intersection exit features of the exit. An aggregated exit encoding based on aggregating all of the intersection exit encodings for the exits is determined. For each of the exits, an exit probability of the exit that the moving obstacle likely exits the intersection through the exit is determined based on the obstacle state evolution and the aggregated exit encoding. Thereafter, a trajectory of the ADV is planned to control the ADV to avoid a collision with the moving obstacle based on the exit probabilities of the exits.

Abstract: An obstacle state evolution of a spatial position of a moving obstacle over a period of time is determined. A lane-obstacle relation evolution of the moving obstacle with each of one or more lanes near the moving obstacle over the period of time is further determined. An intended movement of the moving obstacle is predicted based on the obstacle state evolution and the lane-obstacle evolution. Thereafter, a trajectory of the ADV is planned to control the ADV to avoid a collision with the moving obstacle based on the predicted intended movement of the moving obstacle. The above process is iteratively performed for each of the moving obstacles detected within a predetermined proximity of the ADV.

Abstract: In one embodiment, a process is performed during controlling Autonomous Driving Vehicle (ADV). Microphone signals sense sounds in an environment of the ADV. The microphone signals are combined and filtered to form an audio signal having the sounds sensed in the environment of the ADV. A neural network is applied to the audio signal to detect a presence of an audio signature of an emergency vehicle siren. If the siren is detected, a change in the audio signature to make a determination as to whether the emergency vehicle siren is a) moving towards the ADV, or b) not moving towards the ADV. The ADV can make a driving decision, such as slowing down, stopping, and/or steering to a side, based on if the emergency vehicle siren is moving towards the ADV.

Abstract: In one embodiment, a system generates a plurality of trajectory candidates for an autonomous driving vehicle (ADV) from a starting point to an end point of a particular driving scenario. The system generates a reference trajectory corresponding to the driving scenario based on a current state of the ADV associated with the starting point and an end state of the ADV associated with the end point, where the reference trajectory is associated with an objective. For each of the trajectory candidates, the system compares the trajectory candidate with the reference trajectory to generate an objective cost representing a similarity between the trajectory candidate and the reference trajectory. The system selects one of the trajectory candidates as a target trajectory for driving the ADV based on objective costs of the trajectory candidates.

Abstract: In one embodiment, a process is performed during controlling Autonomous Driving Vehicle (ADV). A plurality of point confidence scores are determined, each defining a reliability of a corresponding point on a trajectory of a moving obstacle. At least one of the point confidence scores is determined based on a) an overall trajectory confidence score, and b) at least one environmental factor of the obstacle. The ADV is controlled based on the trajectory of the moving obstacle and at least one of the plurality of point confidence scores.

Abstract: In one embodiment, an autonomous driving system of an ADV perceives a driving environment surrounding the ADV based on sensor data obtained from various sensors, including detecting one or more lanes and at least a moving obstacle or moving object. For each of the lanes identified, an NN lane feature encoder is applied to the lane information of the lane to extract a set of lane features. For a given moving obstacle, an NN obstacle feature encoder is applied to the obstacle information of the obstacle to extract a set of obstacle features. Thereafter, a lane selection predictive model is applied to the lane features of each lane and the obstacle features of the moving obstacle to predict which of the lanes the moving obstacle intends to select.

Abstract: A method, an apparatus, a storage medium, and an electronic device for testing dynamic parameter of vehicle are provided. The method for testing dynamic parameter of vehicle provided by the present disclosure includes: first obtaining a control parameter for an autonomous vehicle; then controlling the vehicle to travel automatically under a given environment according to the control parameter, detecting and recording traveling data of the vehicle; and at last determining a dynamic parameter of the vehicle according to the traveling data. According to the method for testing dynamic parameter provided by the present disclosure, the characteristic of automatic driving of an autonomous vehicle is utilized to achieve an automatic measurement of the dynamic parameter, thereby reducing cost for calibrating the vehicle and significantly improving safety during the test. Additionally, human error caused by manually driving during the test can be avoided effectively.

Abstract: In one embodiment, in response to detecting an obstacle based on a driving environment surrounding an autonomous driving vehicle (ADV), a system projects the obstacle onto a station-time (ST) graph, where the ST graph indicates a location of the obstacle relative to a current location of the ADV at different points in time. The system determines a first set of end points that are not overlapped with the obstacle within the ST graph, wherein each of the end points in the first set represents a possible end condition. The system generates a first set of trajectory candidates between a starting point representing the current location of the ADV and the end points of the first set based on the ST graph. The system selects one of the trajectory candidates in the first set using a predetermined trajectory selection algorithm to control the ADV in view of the obstacle.

Abstract: A moving object such as a vehicle is identified within an intersection having multiple exits. The moving object and the intersection and its exits may be identified based on sensor data obtained from various sensors mounted on an ADV. An exit coordinate map is generated based on the orientation of the moving object and a relative position of each of the exits of the intersection with respect to the current position of the moving object. For each of the exits, an exit probability of the exit that the moving object likely exits the intersection using the exit coordinate map. Thereafter, a trajectory of the ADV is planned to navigate through the intersection to avoid the collision with the moving object based on the exit probabilities of the exits of the intersection. The above process is iteratively performed for each of the moving objects detected within the proximity of the intersection.

Abstract: A battery pack(1) includes a housing (2) and an array of electrochemical cells (80) disposed in the housing (2). The housing (2) includes a container (3) and a lid (30) that closes an open end of the container (3). The container (3) has a base (4), a sidewall (8) that surrounds the base (4), and a spring plate (110) disposed inside the sidewall (8) between the cells (80) and the sidewall (8). The spring plate (110) is free standing within the container (3) and applies a spring force to the cell array that restrains the cells (80) along an axis normal to the surface of the spring plates (110). The lid (30) includes inwardly-protruding pins (50, 60) that further restrain the cells (80) within the housing (2).

Abstract: In response to perceiving a moving object, one or more possible object paths of the moving object are determined based on the prior movement predictions of the moving object, for example, using a machine-learning model, which may be created based on a large amount of driving statistics of different vehicles. For each of the possible object paths, a set of trajectory candidates is generated based on a set of predetermined accelerations. Each of the trajectory candidates corresponds to one of the predetermined accelerations. A trajectory cost is calculated for each of the trajectory candidates using a predetermined cost function. One of the trajectory candidates having the lowest trajectory cost amongst the trajectory candidates is selected. An ADV path is planned to navigate the ADV to avoid collision with the moving object based on the lowest costs of the possible object paths of the moving object.

Abstract: In one embodiment, a data processing system for an autonomous driving vehicle (ADV) includes a processor, and a memory coupled to the processor to store instructions, which when executed by the processor, cause the processor to perform operations. The operations include generating a station-time (ST) graph based on perception data obtained from one or more sensors of the ADV, the ST graph including representing a location of an obstacle at different points in time, obtaining a tensor based on the ST graph, the tensor including a plurality of layers, the plurality of layers including a first layer having data representing one or more obstacles on a path in which the ADV is moving, applying a machine-learning model to the plurality of layers of the tensor to generate a plurality of numerical values, the plurality of numerical values defining a potential path trajectory of the ADV, and determining a path trajectory of the ADV based on the plurality of numerical values.

Abstract: An ADV may determine a predicted path for a moving obstacle. The ADV may determine a predicted area based on the predicted path. The ADV may determine a path for the ADV based on the predicted area. The path for the ADV may avoid the predicted area when determining the path for the ADV.

Abstract: A travelling track prediction method and device for a vehicle are provided. The method includes that: calculating a plurality of potential positions to which is to be reached by a target vehicle within a preset time within a sensible range of a main vehicle; selecting at least two positions from the plurality of potential positions as target positions; predicting a plurality of travelling tracks to each target position for the target vehicle based on travelling state information of the target vehicle; and selecting at least one travelling track from the plurality of travelling tracks as a prediction result of the target vehicle based on environment information around the target vehicle. According to the embodiments, the travelling track of the target vehicle around the main vehicle may be accurately predicted based on travelling state information and the environment information.

Abstract: In one embodiment, a data processing system for an autonomous driving vehicle (ADV) includes a processor, and a memory coupled to the processor to store instructions, which when executed by the processor, cause the processor to perform operations. The operations include generating a station-time (ST) graph based on perception data obtained from one or more sensors of the ADV, the ST graph including representing a location of an obstacle at different points in time, obtaining a tensor based on the ST graph, the tensor including a plurality of layers, the plurality of layers including a first layer having data representing one or more obstacles on a path in which the ADV is moving, applying a machine-learning model to the plurality of layers of the tensor to generate a plurality of numerical values, the plurality of numerical values defining a potential path trajectory of the ADV, and determining a path trajectory of the ADV based on the plurality of numerical values.