Abstract: Provided is a method and an apparatus for determining a vehicle behavior, and more specifically, to a method and an apparatus for determining a vehicle behavior for bottleneck congestion control in a bottleneck section. Tn apparatus for determining a vehicle behavior may include an information collection unit collecting surrounding information of a target driving vehicle from a road side unit (RSU), a vehicle observation unit obtaining observation information based on the target driving vehicle from a sensing module mounted on the target driving vehicle, a reward determination unit determining a reward for the target driving vehicle through a reward function which uses the surrounding information and the observation information, a model training unit updating and training a decision making model through the reward, and a behavior determination unit determining a behavior of the target driving vehicle by inputting the observation information into the decision making model.

Abstract: A traffic control method and apparatus with an autonomous vehicle based on adaptive reward. A traffic control apparatus for an autonomous vehicle based on adaptive reward comprises an information observation unit that collects observation information from a sensing module of an autonomous vehicle or a roadside unit (RSU); a policy execution unit that decides on an action including adjusting acceleration and changing lanes of the autonomous vehicle based on the observation information and policy; and a reward determination unit that determines reward according to observation information at a next timestep according to the decision made, wherein reward in the reward determination unit includes penalty in an event of an accident and reward when driving, wherein the reward when driving includes an adaptive target speed reward term, a successful lane change reward term, and a safety distance compliance reward term that are adaptively determined according to road traffic.

Abstract: A deep reinforcement learning-based decision-making apparatus through prior data and selective imitation learning is disclosed. The deep reinforcement learning-based decision-making apparatus comprises a prior data collection unit configured to collect prior data from one or more other agents; a prior data processing unit configured to process the collected prior data into data including state, action, next state, and reward; and a policy learning unit configured to learn policy of an ego agent using the processed prior data and interaction data including state, action, next state, and reward obtained through real-time interaction with environment.

Abstract: A traffic control method and apparatus with an autonomous vehicle based on adaptive reward. A traffic control apparatus for an autonomous vehicle based on adaptive reward comprises an information observation unit that collects observation information from a sensing module of an autonomous vehicle or a roadside unit (RSU); a policy execution unit that decides on an action including adjusting acceleration and changing lanes of the autonomous vehicle based on the observation information and policy; and a reward determination unit that determines reward according to observation information at a next timestep according to the decision made, wherein reward in the reward determination unit includes penalty in an event of an accident and reward when driving, wherein the reward when driving includes an adaptive target speed reward term, a successful lane change reward term, and a safety distance compliance reward term that are adaptively determined according to road traffic.

Abstract: Provided is a method and an apparatus for determining a vehicle behavior, and more specifically, to a method and an apparatus for determining a vehicle behavior for bottleneck congestion control in a bottleneck section. Tn apparatus for determining a vehicle behavior may include an information collection unit collecting surrounding information of a target driving vehicle from a road side unit (RSU), a vehicle observation unit obtaining observation information based on the target driving vehicle from a sensing module mounted on the target driving vehicle, a reward determination unit determining a reward for the target driving vehicle through a reward function which uses the surrounding information and the observation information, a model training unit updating and training a decision making model through the reward, and a behavior determination unit determining a behavior of the target driving vehicle by inputting the observation information into the decision making model.