Holistic Range Prediction for Electric Vehicles Stefan Khler, FZI - PowerPoint PPT Presentation

Holistic Range Prediction for Electric Vehicles Stefan Köhler, FZI "apply & innovate 2014" 24.09.2014 S. Köhler, 29.09.2014

Outline Overview: Green Navigation Influences on Electric Range Simulation Toolchain System Integration Summary and Outlook S. Köhler, 29.09.2014 2

Green Navigation: Project Goals Reliable range prediction through Consideration of route, traffic, vehicle parameters, charging stations, weather forecast, driver behavior Adaption of driving strategies and hints for different EV models, load and driver Decentralized and local route calculation providers based on a well-defined interface S. Köhler, 29.09.2014 3

Green Navigation: Project Goals Deterministic characterization of the electric Vehicle vehicle Characterization of driver Driver Behavior Consideration of 3D map data (slopes, curvature, crossings, charging stations) Speed limits Consideration of weather impact (HVAC, wind, humidity, temperature, solar 3D Route Profile radiation, etc.) innovative infrastructure to include real-time Traffic Light Crossings vehicle data and cloud based service providers Weather model based development and early simulation using a novel integration and Dynamic Traffic testing platform S. Köhler, 29.09.2014 4

Green Navigation: Project Content Application Gateway Integration Driver Routing and Education Validation Range Prediction S. Köhler, 29.09.2014 5

Green Navigation: Overview Range Prediction Charging Traffic Flow Static HVAC & Stations Information Consumers Thermal Vehicle Model Parameters 3D Map Data Powertrain Weather Information Navigation Vehicle Energy Environ- Services Model Management ment Model ADAS Energy Consumption Prediction Range Estimation Driver Driver Sensors Identification Model S. Köhler, 29.09.2014 6

Influences: Learning of Driver Influences Average deviation from speed limit Average accelerator pedal velocity Average brake pressure change Average time gap between gas and brake pedal usage 30 Bremsdruckanstieg [bar/s] 25 20 15 4 3 driver 3.5 characteristics Wechselzeit [s] 3 (Clustering) 2.5 2 1 2 1.5 3 1 4 6 8 10 12 14 50 60 70 80 90 100 110 Geschwindigkeitsüberschreitung [km/h] Gaspedalanstieg [%/s] S. Köhler, 29.09.2014 7

Influences: Driver Identification Goal: Identification of Driver Selection and improvement of learned driver model Adaption of driving hints according to drivers‘ preferences Approach Identification via video or depth map image data Parameterization of driver model Automatic serialization/deserializaton of driver model Driver Identification Estimation of head attitude based on depth map and color image Extraction of silhouette from depth map data Identification of driver via SVM Driver specific profile and models S. Köhler, 29.09.2014 8

Influences: Weather Impact Identification of significant weather parameters Temperature, solar radiation Wind velocity and heading Ambient pressure Sensitivity analysis Weather data for target area (Karlsruhe-Stuttgart) Coverage of 14,000 km 2 target area (100 x 140 km) Cloud based service provider Relevant parameters Accurate temporal and spatial resolution forecast well-defined interface S. Köhler, 29.09.2014 9

Simulation Toolchain: Validation- and Test-Environment for EV Validation Office (PC/Notebook) System Experience Platform Mobile Driving Simulator Stationary Vehicle data Test Drives S. Köhler, 29.09.2014 10

Simulation Toolchain: Validation- and Test-Environment for EV Parameterization Visualization Extended Interfaces S. Köhler, 29.09.2014 11

Simulation Toolchain: Vehicle Models static consumer measure.- measure- ment ment modeling Driving / operation modeling strategies simulation simulation Introduction into Co-Simulation Toolchain available component models / parameters S. Köhler, 29.09.2014 12

Simulation Toolchain: Vehicle Parameterization electric vehicle parameters Parameterization (batteries, motor, control units) S. Köhler, 29.09.2014 13

Simulation Toolchain: Vehicle Parameterization S. Köhler, 29.09.2014 14

Simulation Toolchain: Environment IPG CarMaker coupled with Driver model Google Traffic Map Data Weather service provider Temperature profile (over route or time) Humidity and pressure Solar radiation Wind velocity and heading S. Köhler, 29.09.2014 15

System Integration: Architecture Display, Control and Configuration Provisioning of Data via Navi, Android-System PTV, Bosch services standardized communication channel, security and privacy of data guaranteed CarMedialab UMTS (3G) Service Flea- Box Tunnel RP-System Processing of Data services and results are analyzed Onboard Systems Daimler FleetBoard for data acquisition and distribution FZI, CarMedialab S. Köhler, 29.09.2014 16

System Integration: System Experience Platform Integration of all functions in an Human-in-the-Loop demonstrator S. Köhler, 29.09.2014 17

Summary and Outlook Summary Sensitivity analysis weather/ driver Simulation Toolchain components and parameters environment Analysis and abstraction for energy and range prediction Server based (fleet management) Onboard (private transport) Specification of architecture and interfaces Integration in Office-Simulation and System Experience Platform Integration in vehicle modular Future Work Test drives for further evaluation and tuning of functions and models Focus on driver education S. Köhler, 29.09.2014 18

THANK YOU! e-mobil BW GmbH Leuschnerstr. 45 I 70176 Stuttgart Telefon: +49 711 892385-0 Telefax: +49 711 892385-49 info@e-mobilbw.de www.e-mobilbw.de S. Köhler, 29.09.2014