Intelligent Power Dataset

The Intelligent Power Dataset is a multi-modal spatio-temporal data set generated based on the simulation platform of Internet of vehicles, which integrates the vehicle state, surrounding traffic elements and road facilities information, covers multi-dimensional driving scenarios of light/heavy vehicles under Chinese standard driving conditions, and provides high-fidelity benchmark data support for the research of prediction algorithms of intelligent transportation systems.

Multi-source heterogeneous information fusion: multi-dimensional spatio-temporal data such as vehicle state (historical speed), surrounding traffic elements (distance/speed of leading vehicle, distance of other vehicles), traffic facility information (signal light distance) and traffic flow characteristics (lane density) were integrated.
Scenario diversity coverage: Typical driving scenarios of light passenger vehicles and heavy commercial vehicles are constructed based on Chinese standard operating conditions (CLTC/CHTC), covering differentiated driving modes such as low-speed congestion, medium-speed commuting, and high-speed cruising.
High-fidelity simulation modeling: it is realized by SUMO microscopic traffic simulation platform, Krauss car-following model and SL2015 lane-changing model are used as the bottom layer, and Chinese road parameters are combined to generate physically realistic vehicle trajectory data.
Dynamic time series representation ability: The driving cycle is constructed by the splicing method of short travel segments, the sampling frequency is 1Hz, and the time series covers the transient working conditions such as acceleration, deceleration and idle, so as to strengthen the description of the nonlinear characteristics of the actual traffic flow.

Download Links: