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 control command is generated with an MPC controller, the MPC controller including a cost function with weights associated with cost terms of the cost function. The control command is applied to a dynamic model of an autonomous driving vehicle (ADV) to simulate behavior of the ADV. One or more of the weights are based on evaluation of the dynamic model in response to the control command, resulting in an adjusted cost function of the MPC controller. Another control command is generated with the MPC controller having the adjusted cost function. This second control command can be used to effect movement of the ADV.

Abstract: A method of navigating an autonomous driving vehicle (ADV) includes determining a target function for an open space model based on one or more obstacles and map information within a proximity of the ADV, then iteratively performing first and second quadratic programming (QP) optimizations on the target function. Then, generating a second trajectory based on results of the first and second QP optimizations to control the ADV autonomously using the second trajectory. The first QP optimization is based on fixing a first set of variables of the target function. The second QP optimization is based on maximizing a sum of the distances from the ADV to each of the obstacles over a plurality of points of the first trajectory, and minimizing a difference between a target end-state of the ADV and a determined final state of the ADV using the first trajectory.

Abstract: A simulation of an autonomous driving vehicle (ADV) includes capturing first data that includes a control command output by an autonomous vehicle controller of the ADV, and capturing second data that includes the control command being implemented at a control unit of the ADV. The control command, for example, a steering command, a braking command, or a throttle command, is implemented by the ADV to affect movement of the ADV. A latency model is determined based on comparing the first data with the second data, where the latency model defines time delay and/or amplitude difference between the first data and the second data. The latency model is applied in a virtual driving environment.

Abstract: In one embodiment, control of an autonomous driving vehicle (ADV) includes determining a current scenario of the ADV. Based on the scenario, a control algorithm is selected among a plurality of distinct control algorithms as the active control algorithm. One or more control commands are generated using the active control algorithm, based one or more target inputs. The control commands are applied to effect movement of the ADV.

Abstract: In one embodiment, a computer-implemented method of operating an autonomous driving vehicle (ADV) includes perceiving a driving environment surrounding the ADV based on sensor data obtained from one or more sensors mounted on the ADV, determining a driving scenario, in response to a driving decision based on the driving environment, applying a predetermined machine-learning model to data representing the driving environment and the driving scenario to generate a set of one or more driving parameters, and planning a trajectory to navigate the ADV using the set of the driving parameters according to the driving scenario through the driving environment.

Abstract: Systems and methods are disclosed for optimizing values of a set of tunable parameters of an autonomous driving vehicle (ADV). The controllers can be a linear quadratic regular, a “bicycle model,” a model-reference adaptive controller (MRAC) that reduces actuation latency in control subsystems such as steering, braking, and throttle, or other controller (“controllers”). An optimizer selects a set tunable parameters for the controllers. A task distribution system pairs each set of parameters with each of a plurality of simulated driving scenarios, and dispatches a task to the simulator to perform the simulation with the set of parameters. Each simulation is scored. A weighted score is generated from the simulation. The optimizer uses the weighted score as a target objective for a next iteration of the optimizer, for a fixed number of iterations. A physical real-world ADV is navigated using the optimized set of parameters for the controllers in the ADV.

Abstract: In one embodiment, a method of training dynamic models for autonomous driving vehicles includes the operations of receiving a first set of training data from a training data source, the first set of training data representing driving statistics for a first set of features; training a dynamic model based on the first set of training data for the first set of features; determining a second set of features as a subset of the first set of features based on evaluating the dynamic model, each of the second set of features representing a feature whose performance score is below a predetermined threshold. The method further includes the following operations for each of the second set of features: retrieving a second set of training data associated with the corresponding feature of the second set of features, and retraining the dynamic model using the second set of training data.

Abstract: An obstacle is detected based on sensor data obtained from a plurality of sensors of the ADV. A distribution of a plurality of positions of the obstacle at a point of time may be predicted. A range of positions of the plurality of positions of the obstacle may be determined based on a confidence level of the range. A modified shape with a modified length of the obstacle may be determined based on the range of positions of the obstacle. A trajectory of the ADV based on the modified shape with the modified length of the obstacle may be planned. The ADV may be controlled to drive according to the planned trajectory to drive safely to avoid a collision with the obstacle.

Abstract: An obstacle is detected based on sensor data obtained from a plurality of sensors of the ADV. Multiple trajectories of the obstacle are predicted with corresponding probabilities including a first predicted trajectory of the obstacle with a highest probability and a second predicted trajectory of the obstacle with a second highest probability. A cautionary trajectory of the ADV is planned based on at least one of a difference between the highest probability and the second highest probability or a consequence of the second trajectory. The ADV is to drive with a speed lower than a speed limit and prepare to stop in the cautionary trajectory. The ADV is controlled to drive according to the cautionary trajectory.

Abstract: According to various embodiments, described herein is a method of creating a simulation environment with multiple simulation obstacle vehicles, each with a different human-like driving style. Training datasets with different driving styles can be collected from individual human drivers, and can be combined to generate mixed datasets, each mixed dataset including only data of a particular driving style. Multiple learning-based motion planner critics can be trained using the mixed datasets, and can be used to tune multiple motion planners. Each tuned motion planner can have a different human-like driving style, and can be installed in one of multiple simulation obstacle vehicles. The simulation obstacle vehicles with different human-like driving styles can be deployed to the simulation environment to make the simulation environment more resemble a real-world driving environment.

Abstract: According to some embodiments, described herein is a method and a system for guaranteeing safety at a control level of an ADV when at least a portion of a planned path generated by a planning module of the ADV is uncertain due to traffics and/or road condition changes. The planning module, when generating a path, also generate a confidence level of each segment of the path based on one or more of perception data, map information, or traffic rules. The confidence levels are decreasing further away from the ADV. When the control module of the ADV obtains the path and the associated confidence levels, the control module issue control commands to track only one or two segments whose confidence levels exceeds a threshold hold, and issue default control commands for the rest of the path.

Abstract: In one embodiment, a computer-implemented method for optimizing a controller of an autonomous driving vehicle (ADV) includes obtaining several samples, each sample having a set of parameters, iteratively performing, until a predetermined condition is satisfied: determining, for each sample, a score according to a configuration of the controller based on the set of parameters of the sample, applying a machine learning model to the samples and corresponding scores to determine a mean function and a variance function, producing a new sample as a minimum of a function of the mean function and the variance function with respect to an input space of the set of parameters, adding the new sample to the several samples, and outputting the new sample as an optimal sample, where parameters of the optimal sample are utilized to configure the controller to autonomously drive the ADV.

Abstract: In one embodiment, an ADV is routed by executing a first driving scenario that is active. The first driving scenario is one of a plurality of driving scenario types, each driving scenario type being associated with one or more stages to be executed while a corresponding driving scenario type is active. Based on an environmental condition around the ADV, a second driving scenario is set as active. The ADV is routed by executing the second driving scenario. When the second driving scenario exits, execution of the first driving scenario resumes at the one or more stages of the first driving scenario that remains to be executed.

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: Described herein are a method of training a learning-based critic for tuning a rule-based motion planner of an autonomous driving vehicle, a method of tuning a motion planner using an automatic tuning framework that with the learning-based critic. The method includes receiving training data that incudes human driving trajectories and random trajectories derived from the human driving trajectories; training a learning-based critic using the training data; identifying a set of discrepant trajectories by comparing a first set of trajectories, and a second set of trajectories; and refining, at the neural network training platform, the learning-based critic based on the set of discrepant trajectories. The automatic tuning framework can remove human efforts in tedious parameter tuning, reduce tuning time, while retaining the physical and safety constraints of the ruled-based motion planner.

Abstract: In one embodiment, a method determines a route from a first location of an autonomous driving vehicle (ADV) to a second location within an open space, the first location being a current location of the ADV. The method determines an objective function based on the route, the objective function having a set of costs for maneuvering the ADV from the first location to the second location. The method determines environmental conditions of the open space and uses the environmental conditions to determine a set of weights, each weight to be applied to a corresponding cost of the objective function. The method optimizes the objective function in view of one or more constraints, such that an output of the objective function reaches minimum while the one or more constraints are satisfied and generates a path trajectory with the optimized objective function to control the ADV autonomously according to the path trajectory.

Abstract: In one embodiment, a system/method generates a driving trajectory for an autonomous driving vehicle (ADV). The system perceives an environment of an autonomous driving vehicle (ADV). The system determines one or more bounding conditions based on the perceived environment. The system generates a first trajectory using a neural network model, wherein the neural network model is trained to generate a driving trajectory. The system evaluates/determines if the first trajectory satisfies the one or more bounding conditions. If the first trajectory satisfies the one or more bounding conditions, the system controls the ADV autonomously according to the first trajectory. Otherwise, the system controls the ADV autonomously according to a second trajectory, where the second trajectory is generated based on an objective function, where the objective function is determined based on at least the one or more bounding conditions.

Abstract: According to various embodiments, systems, methods, and mediums for operating an autonomous driving vehicles (ADV) are described. The embodiments use a number of machine learning models to extract features individually from audio data and visual data captured by sensors mounted on the ADV, and then to fuse these extracted features to create a concatenated feature vectors. The concatenated feature vector is provided to a multiplayer perceptron (MLP) as input to generate a detection result related to the presence of an emergency vehicle in the surrounding environment. The detection result can be used by the ADV to take appropriate actions to comply with the local traffic rules.

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.