ADAS & ADASIS v2. Sinisa Durekovic, NAVTEQ

Transcription

1 ADAS & ADASIS v2 Sinisa Durekovic, NAVTEQ

2 Interrupts enabled! If you have question, please feel free to interrupt me! 2

3 Who am I? Sinisa Durekovic NAVTEQ, Principal Engineer, Team Lead SPM CE EMEA In-Vehicle Editor ADASIS v2 Protocol Specification Coordinator ADASIS v2 Horizon Reconstructor Implementation Task Force MAPS & ADAS project ADASIS Forum since 2005 interactive, EcoMove, Genivi, Dynamic Pass Predictor, Curve/Speed Warning, Intelligent ACC, Green ACC, Principal Consultant at NAVTEQ since 2003 Responsible for ADASRP since Tel: +49 (6196) Studied Mathematics at University of Zagreb, Croatia In Software Development since 1986 Working on In-Car Navigation Systems since

4 What is ADAS? Advanced Driver Assistance Systems (ADAS) is the application of automotive sensor, computational and/or digital map technologies to create driving experiences that are: safer less stressful more productive more economical 4

5 ADAS Application: Adaptive Cruise Control Standard Cruise Control Driver sets desired driving speed Adaptive Cruise Control Using radars/lidars, adjust speed to the vehicle in front Accelerate to desired speed if lane is clear 5

6 ADAS Application: Adaptive Cruise Control 6

7 ADAS Application: Adaptive Cruise Control ACC Issue: Wrong detection of car in front at curve! 7

8 ADAS Application: Adaptive Cruise Control ACC Issue: Acceleration on Highway Exits! 8

9 ADAS Application: Adaptive Cruise Control ACC Issue: Excessive curve/crossing speed! 9

10 ADAS Application: Adaptive Cruise Control ACC Issue: Wrong detection of car in front at curve Solution approach: Use Digital Map identify such situation and refine locking algorithm ACC Issue: Acceleration on Highway exits Solution: Use Digital Map to identify when car will leave the highway ACC Issue: Excessive Curve/Crossing Speed Solution: Use Digital Map to detect sharp curves ahead, calculate maximum speed Solution: Reduce speed or stop vehicle at crossings ACC Issue: Set speed stays set unless driver change it Solution: Identify Legal Speed Limit ahead; adjust ACC top speed according to the speed limit Intelligent Adaptive Cruise Control 10

11 Evolution of the ACC Cruise Control + radar Adaptive Cruise Control + map with ADAS attributes Intelligent Cruise Control + fuel optimization algorithms Green Cruise Control 11

12 Use of digital maps with ACC BMW platform Application was launched in March 2005 in BMW 3, 5, 6 series in Europe Digital map integrated with Adaptive Cruise Control Maps used to adjust acceleration into three classes, high middle and low acceleration. 12

13 Freightliner: Run Smart Predictive Cruise 13

14 Freightliner: Run Smart Predictive Cruise Precision Instruments Predicts the road ahead approximately one mile in front of vehicle s Immediate location Driver-Free Controls Invisible to the driver functions with no distractions or on-screen messages Same driver interaction as standard cruise control Works seamlessly with standard cruise control 14

15 Fuel Economy and Emissions Reduction fuel consumption [l/100km] trip time [min] % % number of gearshifts [-] % LC CC LC CC LC CC 15

16 BMW Dynamic Pass Predictor 16

17 Adaptive Front Lighting Front lights may project beyond the road, rather than into the curve Wheel sensor-based implementation points lights along the current tangent to the curve which is not sufficient lead Control of beam elevation on slopes Control of beam width and range in urban or rural Conventional Map-based Map-Based Conventional Map-Based Conventional 17

18 Map-Enabled and Map-Enhanced ADAS Applications Non-map ADAS Applications Do not need and can t use digital map Example: Park Distance Control (PDC) Map-Enhanced ADAS Application Can function without, but work better with the digital map Example: (Intelligent) ACC Map-Enabled ADAS Application Only possible with use of digital map Example: Curve Speed Warning In any given moment, only part of the Digital Map is required for the ADAS function especially most probable path ahead Electronic Horizon 18

19 Map Supported ADAS Positioning and usage based Applications Pay as you Drive Insurance Road User Charging Congestion Charging Telematics, E- / B-Call Traffic Probing / Fleet Management Driver Warning Active Safety Driver Warnings & Alerts Curve Speed Warning Lane Departure Warning Overtaking Assistant Drowsiness Monitor 19 Vehicle Efficiency Green Driving Assistant Green ACC Smart Transmission Hybrid Power Management Intelligent Cruise Control Predictive Frontlight Sys Roll-over alert Lane Keeping Assistant Electronic Stability Control Adaptive Brake Systems

20 ADAS Digital Map is a special Sensor The ADAS digital map provides important road-based information on the Electronic Horizon which is combined with other sensors or may be used independently Radar Torque Video Gyro Lidar NAVTEQ Maps GPS 20

21 What is an Electronic Horizon? 1. Find current position 2. Find all possible paths 3. Calculate probabilities Electronic Horizon 4. Starting from current position, follow highest probable path Most Probable Path 21

22 Electronic Horizon Data structure containing: Vehicle position location with reference to the map and vehicle dynamics information. Road network in front of vehicle including attributes all potential driving paths from the current position up to a certain depth. Probabilites the driving probability of each path Route Information optional route information to influence the driving probabilities dynamically. Other names: ADAS Horizon ADASIS Horizon Extended Driver Horizon ehorizon 22

23 Organization ADAS Interface Specification (ADASIS) Forum Initiated by Navteq in May 2001 Goals: - Define an open standardized data model and structure to represent map data in the vicinity of the vehicle position (i.e. the ADAS Horizon), in which map data are delivered by navigation and/or by a general map data server - Define open standardized API(s) to enable ADAS applications to access the electronic horizon and position-related data of the vehicle m A.D.C. GmbH, AISIN AW, Alpine, Autoliv, Blaupunkt, BMW, DaimlerChrysler, Denso, Ford Research Center Aachen, Hella, Honda, Intermap Technologies, Michael Sena Consulting AB, Mitsubishi Electric Automotive Europe, Navigon, Navteq, Nissan, Opel, PSA, Renault, Robert Bosch, Scania, Siemens VDO Automotive, Tele Atlas, Toyota Motor Corporation, Visteon, Volkswagen, Volvo Technology Corporation (VTEC), Volvo Car Corporation, Zenrin Europe BV 23

24 Projects EU PReVENT/Maps & ADAS project ( ) Develop, test and validate both: - An applicable standard for gathering, certifying, maintaining and providing safety content enhanced digital map databases to be used in ADAS and Navigation Applications - An applicable standard for an interface between ADAS and the in-vehicle map data sources (not necessarily navigation systems) for accessing map data around the vehicle position Blaupunkt, BMW, Bosch, Daimler, Ertico, Ford, IVH, Siemens, Navigon, Navteq, Volvo, Transver, Teleatlas Result: ADASIS v1 Protocol & API ADASIS v2 Initiative ( ) ADASIS v1 too complex Goal: To create simpler, more acceptable protocol Navteq (coordinator), Bosch, BMW, Continental, Daimler, Ford, Opel, PSA, TeleAtlas ADASIS v2 Horizon Reconstructor C-Implementation Task Force (Av2HR-C ITF; ) Goal: To develop Av2HR as Proof-of-concept Result: ADASIS v2 Protocol & API (April 2010), Av2HR 24

25 Architecture of ADAS system (e.g. AFL) AFL module is located on the Head Unit AFL algorithm directly access EH via proprietary EH API Since EH API is OEM-specific, AFL module need to be adopted for each particular implementation of Head-Unit Multiple EH-based application must be placed on the same Head Unit 25

26 Architecture of ADAS system (e.g. AFL) AFL module is located on the dedicated ECU Data required for AFL algorithm are extracted from EH by AFL EH Provider AFL EH Provider sends necessary data over (CAN) bus using AFL-specific protocol Minimum amount of data for the AFL on the bus AFL module takes and uses data directly from the bus Since EH API is OEM-specific, having AFL module on another platform is reduced to re-implementation of the EH AFL Provider, but not whole modules Multiple EH applications must have its own provider and protocols 26

27 Architecture of ADAS system (e.g. AFL) AFL module is located on the dedicated ECU Head Unit sends generic Electronic Horizon data on the bus using defined protocol BMW ADAS Audi PSD ADASIS v2 Daimler, Ford, Opel/GM, Genivi (?) EH Reconstructor on ECU reconstructs copy of the EH on client side; apps can use defined EH API to access the data AFL algorithms access EH data via EH API Multiple EH applications use same Reconstructor, API and data stream 27

28 ADASIS v2 ADASIS v2 ADASIS v2 ADASIS v2 Reference Architecture 1. ADASIS v2 Horizon Provider (Av2HP) -- maintains the ADASIS v2 Horizon; ADAS Horizon Provider 2. ADASIS v2 Protocol (Av2) -- defines how the ADASIS v2 Horizon will be sent from the Av2HP to the ADAS Applications; 3. ADAS Application -- a client application that receives the ADASIS v2 Protocol messages then reconstructs and uses the ADASIS Horizon; ADASIS v2 Reconstructor -- a common component of ADAS Applications that is built in accordance with this general architecture. The task of the ADASIS v2 Reconstructor is to receive, parse and interpret ADASIS v2 Protocol messages, and, in effect, reconstruct a copy of the ADASIS v2 Horizon on the client side. ADASIS v2 Reconstructor ADASIS v2 CAN Bus ADASIS v2 Reconstructor ADASIS v2 Reconstructor ADAS Application 1 ADAS Application 2 ADAS Application 3 28

29 ADASIS v2 in Action 29

30 Skip ADASIS v2 Protocol 30

31 Categories of the Electronic Horizons Pure Network ADASIS Digital Path Map v2 Representation Horizon Optimized Path Representation 31

32 Types of ADASIS v2 Horizons Stub Stub Path 2 Path 4 Stub Path 1 Stub Path 3 Path 5 Full Horizon Stub Stub Path 2 Path 4 Path 1 Stub Path 3 Single Path with sub-paths Stub Stub Path 2 Stub Single Path with stubs Path 2 Single Path 32

33 Positioning of entities on ADASIS v2 Horizon Position = Path Identifier + Offset 33

34 ADASIS v2 Horizon Building Blocks Path entity container Possible trajectory of the vehicle Uniquely identified by Path Identifier Entities positioned on Av2 Horizon using Path Identifier and Offset STUB denotes start of the path; contains parent Path Identifier and Offset on the parent path where new path starts POSITION position of the vehicle (Path Id+Offset) Profile attribute of the path Traffic Sign, Number of Lanes, Curvatures, Location of attribute defined by Path Identifier and Offset PROFILE SHORT attribute values can fit to 10 bits PROFILE LONG attribute values can fit to 32 bits SEGMENT encapsulates most important attributes for a part of path 34 META-DATA auxiliry information about the Av2 Horizon Speed Limit Units, Driving Side,

35 ADASIS v2 Protocol 6 message types for CAN protocol (but can be implemented on other buses too) POSITION message specifies the current position(s) of the vehicle - Up to 4 position candidates supported STUB message indicates the start of a new path that has origin at the existing one - Up to 58 paths can be maintained in protocol concurrently SEGMENT message specifies the most important profiles of a part of the path - 12 profiles PROFILE SHORT message describes attribute of the path whose value can be expressed in 10 bits - 15 standard, 16 system-specific profiles ; arbitrary interpolations supported PROFILE LONG message describes attribute of the path whose value can be expressed in 32 bits standard, 16 system-specific profiles; arbitrary interpolations supported META-DATA Message contains utility data. Each message can, but not need to, have separate CAN identifier Need at least 1 CAN Id, up to 66 (POSITON+STUB+SEGMENT+META_DATA+62*PROFILE SHORT/LONG) Allows different multiplexing configurations, including AUTOSAR conformance Designed for 10 CAN frames/sec; need minimum 2 CAN frames/sec; no upper limit 35

36 Example: PROFILE SHORT message Always 4 Type of the PROFILE Cyclic counter for error detection Update flag Offset along the path where profile has the first value (left boundary offset) Path index where profile is located First profile value (left boundary value) Distance from left boundary offset to the second value offset (right boundary relative offset) Second profile value (right boundary value) 36

37 v [m/s] Path Attributes Profiles Values of single link attribute change along the path forming path profile 45 PROFILE SHORT Type=Speed 40 Limit Path=? Offset=0 Value= PROFILE SHORT Type=Speed Limit Path=? Offset=100 Value=50 PROFILE SHORT Type=Speed Limit Path=? Offset=200 Value= s [m] Speed Profile 37

38 Path Profiles Bandwidth or Robustness Bandwidth: 2 messages for 3 intervals Robustness: 3 messages for 3 intervals 38

39 Path Profiles - Interpolations Step Interpolation Linear interpolation By adding control points to the profile interval, any other interpolation type is supported. For instance, second-degree polynomial interpolation. 39

40 Standardized profiles SEGMENT message Functional Class Form of Way Effective Speed Limit/Type Number of lanes (both directions) Tunnel Bridge Divided Road Built-up Area Complex Intersection Part of calculated Route PROFILE SHORT (10-bit values) Curvature/Radii Route Number Slope Road Accessibility Road Condition Variable Speed Sign Position Heading Change PROFILE LONG(32-bit values) Longitude Latitude Reserved 7 PROFILE SHORT and 13 LONG types Allows 16 system-specific PROFILE SHORT types and 16 LONG types 40

41 Reconstructing the Horizon: SEGMENTs & STUBs

42 Reconstructing the Horizon: SEGMENTs & STUBs

43 Reconstructing the Horizon: SEGMENTs & STUBs

44 Reconstructing the Horizon: SEGMENTs & STUBs

45 Reconstructing the Horizon: SEGMENTs & STUBs

46 Reconstructing the Horizon: SEGMENTs & STUBs

47 Reconstructing the Horizon: SEGMENTs & STUBs

48 Reconstructing the Horizon: SEGMENTs & STUBs

49 Reconstructing the Horizon: SEGMENTs & STUBs

50 Reconstructing the Horizon: SEGMENTs & STUBs

51 Reconstructing the Horizon: SEGMENTs & STUBs

52 ADASIS v2 ADASIS v2 ADASIS v2 ADASIS v2 Horizon Reconstructor ADASIS version 2 Horizon Reconstructor (MISRA-C implementation) was developed as common effort of the several companies that formed Implementation Task Force (Av2HR-C ITF). ADAS Horizon Provider ADASIS v2 AV2HR-C ITF Goals CAN Bus Refine ADASIS v2 Specification Develop proof-of-concept Av2HR that can be used as base for commercial implementations ADASIS v2 Reconstructor ADASIS v2 Reconstructor ADASIS v2 Reconstructor ADASIS v2 API is public ADAS Application 1 ADAS Application 2 ADAS Application 3 Source code available to ADASIS Forum Members; including MatLab/Simulink support for dspace and xpc systems In scope of the project, several Av2 Horizon Providers were developed (e.g. NAVTEQ ADASRP Av2 Plug-Ins) 52

53 Two Modules and Three APIs 1. Message Codec Decodes messages and takes care about path index/path id conversion and offset conversions 2. Data Store Reconstructs and keeps track of ADASIS v2 Horizon 1. Message Codec API (MC-API) 2. Event API (Ev-API) 3. Data Store API (DS-API) 53

54 Data Store API Application access to ADASIS v2 Horizon 4 main functions 7 meta-data functions 1 auxiliary function 54

55 Event API Message Codec notifies application about new building block of ADASIS v2 Horizon (Vehicle Position, Stub, Profile, ) 5+1 function 55

56 Message Codec API Communication interface notifies Message Codec about arrival of new CAN frame 1 function 56

57 DS-API Example 57

58 ADASIS v2 in Genivi System shall expose the following interfaces: ADASIS v2 Event API Base for ADASIS v2 Protocol Generator for CAN Base for ADASIS v2 Data Store that will provide API for embedded ADAS applications ADASIS v2 Datastore API To be used by embedded ADAS applications ADASIS v2 CAN Protocol To be used by ADASIS client applications 58

59 ADASIS v2 in Genivi 59

60 ADASIS v2 in Genivi But why!? Map-enabled/map-enhanced ADAS slowly but surely finds its place in modern cars BMW 7: Speed Limit Warning Audi A8: Adaptive Front Lighting Future developments: Ghost Driver Warning, Curve Speed Warning, Intelligent ACC, Green Driving, ADAS applications will play even more important role in HEV and FEV, where every Watt counts Insurance companies discovered ADAS-type of applications (Pay-as-you drive) ADAS as extension of Navigation System most important differentiator between in-dash and PND and other mobile devices 60

61 61