Download - ADASIS v2 Introduction
Sinisa Durekovic, NAVTEQ
ADAS & ADASIS v2
Interrupts enabled!
2
If you have question, please feel free to interrupt me!
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 2005-2007
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 2004
E-mail: [email protected]
Tel: +49 (6196) 589-313
Studied Mathematics at University of Zagreb, Croatia
In Software Development since 1986
Working on In-Car Navigation Systems since 19953
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
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
ADAS Application: Adaptive Cruise Control
6
ADAS Application: Adaptive Cruise Control
7
ACC Issue: Wrong detection of car in front at curve!
ADAS Application: Adaptive Cruise Control
8
ACC Issue: Acceleration on Highway Exits!
ADAS Application: Adaptive Cruise Control
9
ACC Issue: Excessive curve/crossing speed!
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
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.
Freightliner: Run Smart Predictive Cruise
13
Freightliner: Run Smart Predictive Cruise
14
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
15
Fuel Economy and Emissions Reduction
35.72 36.73
LC CC
84.99 85.01
LC CC
8
14
LC CC
fuel consumption [l/100km]
- 2.76 %
trip time [min]
- 0.03 %
number of gearshifts [-]
- 42.9 %
16
BMW Dynamic Pass Predictor
17
Adaptive Front LightingFront 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
Conventional
Map-Based Map-Based
Conventional
Map-based
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
Map Supported ADAS
Driver Warning Active Safety
Intelligent Cruise Control
Predictive Frontlight Sys
Roll-over alert
Lane Keeping Assistant
Electronic Stability Control
Adaptive Brake Systems
Driver Warnings & Alerts
Curve Speed Warning
Lane Departure Warning
Overtaking Assistant
Drowsiness Monitor
Vehicle Efficiency
Green Driving Assistant
Green ACC
Smart Transmission
Hybrid Power
Management
Positioning and usage based Applications
19
Pay as you Drive InsuranceRoad User ChargingCongestion ChargingTelematics, E- / B-CallTraffic Probing / Fleet Management
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
20
Radar
Gyro
NAVTEQ Maps
GPS
Video
Lidar
Torque
What is an Electronic Horizon?1. Find current position
2. Find all possible paths
3. Calculate probabilities
4. Starting from current position,
follow highest probable path
21
Electronic Horizon
Most Probable Path
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
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
http://www.ertico.com/en/activities/safemobility/adasis_forum.htm#www.ertico.com/adasisforum
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
ProjectsEU PReVENT/Maps & ADAS project (2003-2007)
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 (2008-2010)
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; 2009-2010)
Goal: To develop Av2HR as Proof-of-concept
Result: ADASIS v2 Protocol & API (April 2010), Av2HR
24
Architecture of ADAS system (e.g. AFL)
25
•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
Architecture of ADAS system (e.g. AFL)
26
•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
Architecture of ADAS system (e.g. AFL)
27
•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
ADASIS v2 Reference Architecture
28
ADAS Horizon Provider
AD
AS
IS v
2
CAN Bus
AD
AS
IS v
2
ADAS
Application 1
ADASIS v2
Reconstructor
AD
AS
IS v
2
ADAS
Application 2
ADASIS v2
Reconstructor
AD
AS
IS v
2
ADAS
Application 3
ADASIS v2
Reconstructor
1. ADASIS v2 Horizon Provider (Av2HP) --
maintains the ADASIS v2 Horizon;
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 in Action
29
30
ADASIS v2 Protocol
Skip
Categories of the Electronic Horizons
31
ADASIS v2 Horizon
Optimized Path Representation
Network RepresentationPure Path RepresentationDigital Map
Types of ADASIS v2 Horizons
32
Path 2
Path 2Stub
Stub
Stub
Path 2
Path 1
Path 3
Path 4
Stub
Stub
Stub
Path 2
Path 1
Path 3
Path 4
Stub
Stub
Stub
Path 5Stub
Full Horizon
Single Path
with sub-paths
Single Path
with stubs
Single Path
Positioning of entities on ADASIS v2 Horizon
33
Position = Path Identifier + Offset
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
META-DATA – auxiliry information about the Av2 Horizon
Speed Limit Units, Driving Side, …
34
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.
- 15 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
Example: PROFILE SHORT message
36
Always 4 Type of the PROFILE
Path index where profile is located
Offset along the path where profile has the first value
(left boundary offset)
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)
Cyclic counter for error detection Update flag
Path Attributes Profiles
Values of single link attribute change along the path forming path profile
37
0
5
10
15
20
25
30
35
40
45
0 25 50 75 100 125 150 175 200 225 250 275 300
s [m]
v [m
/s]
Speed Profile
PROFILE SHORT
Type=Speed Limit
Path=?
Offset=0
Value=100
PROFILE SHORT
Type=Speed Limit
Path=?
Offset=100
Value=50
PROFILE SHORT
Type=Speed Limit
Path=?
Offset=200
Value=100
Path Profiles – Bandwidth or Robustness
38
Bandwidth:
2 messages for 3 intervals
Robustness:
3 messages for 3 intervals
Path Profiles - Interpolations
39
Step Interpolation Linear interpolation
By adding control points to the profile
interval, any other interpolation type
is supported.
For instance, second-degree
polynomial interpolation.
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
40
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
Reconstructing the Horizon: SEGMENTs & STUBs
Reconstructing the Horizon: SEGMENTs & STUBs
Reconstructing the Horizon: SEGMENTs & STUBs
Reconstructing the Horizon: SEGMENTs & STUBs
Reconstructing the Horizon: SEGMENTs & STUBs
Reconstructing the Horizon: SEGMENTs & STUBs
Reconstructing the Horizon: SEGMENTs & STUBs
Reconstructing the Horizon: SEGMENTs & STUBs
Reconstructing the Horizon: SEGMENTs & STUBs
Reconstructing the Horizon: SEGMENTs & STUBs
Reconstructing the Horizon: SEGMENTs & STUBs
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).
AV2HR-C ITF Goals
Refine ADASIS v2 Specification
Develop proof-of-concept Av2HR that can be used as base for
commercial implementations
ADASIS v2 API is public
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
ADAS Horizon Provider
AD
AS
IS v
2
CAN Bus
AD
AS
IS v
2
ADAS
Application 1
ADASIS v2
Reconstructor
AD
AS
IS v
2
ADAS
Application 2
ADASIS v2
Reconstructor
AD
AS
IS v
2
ADAS
Application 3
ADASIS v2
Reconstructor
Two Modules and Three APIs
53
1. Message Codec API (MC-API)
2. Event API (Ev-API)
3. Data Store API (DS-API)
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
Data Store API
54
• Application access to ADASIS v2
Horizon
• 4 main functions
• 7 meta-data functions
• 1 auxiliary function
Event API
55
• Message Codec notifies application
about new building block of
ADASIS v2 Horizon (Vehicle
Position, Stub, Profile, …)
• 5+1 function
Message Codec API
56
• Communication interface notifies
Message Codec about arrival of
new CAN frame
• 1 function
DS-API Example
57
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
ADASIS v2 in Genivi
59
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