general assembly 2007-01-30, orbassano (crf) 1 ! safeprobe safeprobe project status of sp1 christian...
TRANSCRIPT
1General Assembly2007-01-30, Orbassano (CRF)
!
SAFEPROBESAFEPROBEProject Status of SP1Project Status of SP1
Christian Zott
Bosch (Schwieberdingen, Germany) [email protected]
SP1 “In-Vehicle Platform and Sensing” leader
2General Assembly2007-01-30, Orbassano (CRF)
Vorgangsname
WP 1.1 Technical Management
Kick Off
M 1.1.1 SAFEPROBE MTR
M 1.1.2 SAFEPROBE Final Review
WP 1.2 Requirements
T 1.2.1 Vehicle Probe Use Cases
T 1.2.2 System Analysis
T 1.2.3 Requirements
WP 1.3 Specification
T 1.3.1 Internal Data Fusion Spec
T 1.3.2 Environment Data Acquisition Spec
T1.3.2 Cooperative Data Fusion
T 1.3.4 Data Refinement Spec
T 1.3.5 HW Platform Spec
T 1.3.6 SW Platform
WP 1.4 Implementation
T 1.4.1 In-Vehicle HW Platform Dev
T 1.4.2 Data Fusion Algo Dev & Impl
T 1.4.3 In-Vehicle SW Platform Dev
T 1.4.4 Data Communication Impl
T 1.4.5 Cooperative On-board Sensors
T 1.4.6 Development of Probe Vehicles
WP 1.5 Test and Validation
T 1.5.1 Validation Plan
T 1.5.2 In lab Test
T 1.5.3 In-Vehicle Platform Validation
T 1.5.4 Performance Analysis
01.02.
31.07.
30.01.
01 02 03 04 05 06 07 08 09 10 11 12 01 02 03 04 05 06 07 08 09 10 11 12 01 02 03 04 05 06 07 08 09 10 11 12 01 02 03 042006 2007 2008 2009
SP1 Schedule and Milestones
23.4. DF spec to peer review (CRF)
24.8. public HW/SW spec to peer rev. (CRF)
24.8. public DF spec to peer review (CRF)
24.7. HW/SW spec to peer review (Bosch)
Activity
4 regular SP1 meetings in 2006: Berlin, Turin, Birmingham, Turin + ~ participation at 6 other SPs + KO + CG
D1.2.1, 1.2.2, 1.2.3 completed
3General Assembly2007-01-30, Orbassano (CRF)
SP1: Partners’ Effort and Involvement
0
10
20
30
40
50
60
70
80
CR
F
RE
GIE
NO
VO
LV
O
PIA
GG
IO
MM
SE
BO
SC
H
SV
DO
IBE
O
ICC
S
MIR
A
WP 1.1 WP 1.2 WP 1.3 WP 1.4 WP 1.5
MM
4General Assembly2007-01-30, Orbassano (CRF)
D1.2.1 Vehicle Probe Use Cases
• SP1 system boundaries and actors defined
• Bottom-up methodology:
Use case creation, merging and selection (voting) based on partners’ expertise
• Tabular use case description →
• Use case elements:
- Scenario attributes- Event types - Event handling
5General Assembly2007-01-30, Orbassano (CRF)
Examples of SP1 Scenarios
6General Assembly2007-01-30, Orbassano (CRF)
D1.2.2 System Analysis of In-Vehicle Data
• Collection of ~ 200 data titles in 6 cluster
- ENP Environmental Perception - NAV Navigation and Positioning- VDM Vehicle Dynamics Management - BOD Body Management - OSS Occupant Safety Systems - SVD Static Vehicle Data
• Allocation of data (cluster) to event types and use cases
• Performance definitions and assessments:
- Time-to-notification (TTN), accuracy & integrity of notification- Data availability trees (scenario & technology dependency)- Typical performance figures- Performance risks for use cases
7General Assembly2007-01-30, Orbassano (CRF)
D1.2.2, In-vehicle Data Examples
Vehicle Dynamics Management
Body Systems Environmental Perception, e.g. Radar / Laserscanner
Occupant Safety Systems
Navigation Systems, e.g. static map data
Vehicle speed Turn signal on Longitudinal distance (range)
Airbags fired Latitude / Longitude of shape points
Wheel speeds Hazard warning lights
Lateral distance or azimuth and elevation
Crash signal status
(no crash / pretensioners / first level / second level)
Curvature of road element
Yaw rate Wiper setting Relative velocity(range rate)
Roll-over detected Speed restriction
Lateral acceleration Low / high beam headlighs on
Relative acceleration Seat belt locked Number of lanes
Longitudinal acceleration
Fog lights on Track score Lane category
ABS/ESP brake intervention
Hood open Object class (e.g. pedestrian, PTW, car, truck, bike)
Lane type (exit, entrance, overtaking, emergency,…)
Brake pedal touched
Outside air temperature
Lane divider type (dashed, solid,…)
Brake light on Ignition on Lane direction (ahead, left, right, u-turn,…)
Friction coefficient estimation
PTW tilted
8General Assembly2007-01-30, Orbassano (CRF)
Qualitative Performance Assessment
Received signalpower > threshold:
reflector infield-of-view / beam
Radar Tracksavailable with
sufficientperformance
Strong dependenceon driving scenario /
event
Enough detections /limited misseddetection rate
Limited falsedetection rate
Appropriate noiseenvironment/
limited interference
Sufficient reflectorresolution in range,azimuth, elevation
Observation to trackassociation with
sufficientperformance
Sufficient transmitsignal power
Sufficient scan rate,e.g. pulse repitition
rate
AND
AND
AND
Strong dependenceon technology
Strong dependenceon environmental
conditions
No occlusion byother vehicles/
obstacles
No occlusion bybuildings and other
road-sideconstructions
Appropriate aspectangle given a
specular reflection
Sufficient reflectorradar-cross-section
Sufficient rx/txantenna gain
for reflector range,azimuth, elevation
Limited clutter /unwanted detections
(e.g. pavement,guardrails,…)
Appropriatedynamical models forreflector movement
(track prediction)
Radar just taken as an example here
9General Assembly2007-01-30, Orbassano (CRF)
D1.2.3 Requirements
Requirement Types
- General- Hardware- Software & Architecture- Sensor- Data Fusion- Use Case derived
Requirement’s Attributes
- System requirement ID- Name- System req’t definition- System req’t relevance- Responsibility- Type of system requirement- Rationale
Requirements
ScenariosUser NeedsUse Cases
Common Basic Scenarios SP4-SP5
User (of the Subsystem) Needs
(SP1-2-3-4-5)
Accident data Analysis +SoA Analysis SP4-SP5Accident data Analysis
+SoA Analysis SP4-SP5
Technology analysis SP1-SP2-SP3Technology analysis
SP1-SP2-SP3
User\ActorsIdentificationSP1-2-3-4-5
User\ActorsIdentificationSP1-2-3-4-5
Use Cases (SP1-2-3-4-5)
High Level User NeedsGuidelines
(SP7)Guidelines(SP7)
High Level StakeholdersIdentification
(SP7)
High Level StakeholdersIdentification
(SP7)
Tech Scenarios SP1-SP2-SP3
(SP1-5)
System Requirements SP3
System Requirements SP5
System Requirements SP2
System Requirements SP4
System Requirements SP1
High Level Description
10General Assembly2007-01-30, Orbassano (CRF)
WP1.3 Specification Results
Current Architecture Clients:• SP4/5 applications • SafeSpot message generation• Visualization of current LDM contents
VANET: Vehicular Ad-hoc NetworkLDM: Local Dynamic MapQ-API: Query APIT-API: Transaction APISP4/5
cmp Component Model
Certified Client #1,2,...,M
Applicatioin #N Broker
Application #N (SP4/5)
Certified Client #N+1Broker management
Sensor #N (Radar/Laserscanner) VANET
communication
Data Fusion with Database Surv eillance
#1 #N commap #N+1 #N+x
Transaction Query
Transaction Call Generation
Ego Postioning (v ehicle only)
Object refinement Situation refinement
LDM Database (SP1/2) Serv er
Not Certified Clients #M+1,..., N
Application #1 (SP4/5)
Message Generation & Management
VANET communication
SP3: LDM Margin Generation (v ehicle
only)
Message Broker
Broker
Static Map
Sensor #1 (GPS)
HMI Management (SMA)Application #1 Broker
Nav igation Data
Traffic Control States (infrastructure only)
HMI subsystem (v ehicle only)
Message Stack
Q-API
Q-API
Q-API
T-API
Q-API
LDM data
appl #1dataselection
SP3
Brokers:• cast requests
into queries and subscriptions to notifications
• select/filter data• translate messages,
notifications• remote procedure
calls over platform network
SP1-5,7
11General Assembly2007-01-30, Orbassano (CRF)
WP 1.3 Specification Results
Current Architecture (cont’d) Data Fusion on Multiple Levels:
• Raw sensor data, e.g. ranges, angles
• Features, e.g. edges, reflection points
• Object data, e.g. bounding boxes, lane lines, object attributes
• Situations/Events, e.g. relations between objects, trajectory intersections, region/zone entering or exiting
cmp Component Model
Certified Client #1,2,...,M
Applicatioin #N Broker
Application #N (SP4/5)
Certified Client #N+1Broker management
Sensor #N (Radar/Laserscanner) VANET
communication
Data Fusion with Database Surv eillance
#1 #N commap #N+1 #N+x
Transaction Query
Transaction Call Generation
Ego Postioning (v ehicle only)
Object refinement Situation refinement
LDM Database (SP1/2) Serv er
Not Certified Clients #M+1,..., N
Application #1 (SP4/5)
Message Generation & Management
VANET communication
SP3: LDM Margin Generation (v ehicle
only)
Message Broker
Broker
Static Map
Sensor #1 (GPS)
HMI Management (SMA)Application #1 Broker
Nav igation Data
Traffic Control States (infrastructure only)
HMI subsystem (v ehicle only)
Message Stack
Q-API
Q-API
Q-API
T-API
Q-API
LDM data
appl #1dataselection
SP3SP1/2 SP1-5,7
12General Assembly2007-01-30, Orbassano (CRF)
WP 1.3 Specification
Rationale for Client-Server and Event Processing by Platforms
• Development Organisation
- Clear responsibility allocation for application and platform SPs- Avoid duplication of function development by application SPs
• Performance
- Critical functions concentrated in platforms, e.g. latency reduction for relationship searching by efficient geo-data storage (e.g. spatial indexing)
- All applications gain from advances in esp. situation/event level data fusion and geo-data algorithms
- Reduction of traffic on LDM-application interface, e.g. using event driven notifications
• Business models
- Event and fusion abstraction attractive for suppliers of maps, sensors, platforms- Vehicle manufacturers and traffic control get standardized APIs
13General Assembly2007-01-30, Orbassano (CRF)
WP 1.3 Specification Next Steps
• Detailed scenario definition and simulation:
- Vehicles’ trajectories, collisions,… - Sensor & communication configuration: Placement, fields-of-view,...- Resulting detections times (sensing & communication), occlusions,…
• Design of cooperative concepts:
- Sensor-level to central-level fusion: Activity diagrams, data flow- Data models and dictionary, common with SP2, CVIS
Syntax (XML, ASN.1) Semantics: Structure and behaviour (UML, XML)
• Performance prediction of data fusion approaches:
- Timing for sensing, communication, queries, fusion, LDM transaction,…- Accuracy of LDM contents for assumed configurations- “Do-able” scenarios (for given HW: vehicles, equipment)
14General Assembly2007-01-30, Orbassano (CRF)
WP 1.3 Specification
• Open Java Micro Traffic Simulator for scenario exchange and detailed analysis, e.g. generation of timing diagrams
• Executable, doc and sources at SSCA SP1 - SAFEPROBE / Working Area / WP3 / detection scenario simulation
15General Assembly2007-01-30, Orbassano (CRF)
WP 1.3 Specification Next Steps
Some Open Issues
• Detailed contents of LDM (first class diagrams available)
• Only presence or also future in LDM, e.g. predicted trajectory hypotheses?
• What other kinds of “elementary” events in LDM?
• SF Message definition and generation
- Common task for SP1-5, 7
- Until now some types identified:- Beaconing data (periodic, e.g. IDs, position, velocity, heading,…)- Raw data (object or sensor data)- Events (elementary or similar to TPEG-TEC)
- Framing structure likely (c.f. ISO Probe or TPEG-TEC)
- Some core data elements agreed (time & location stamp,…)
16General Assembly2007-01-30, Orbassano (CRF)
WP 1.3 Specification Next Steps
Objects’ Current State :• Moving vehicles’ state paramater:
position, speed, accel., heading,…
• Static road map geometry with attributes (lane direction, connectivity, type…)
• Traffic lights’ phases (e.g. s to change)
• Road/lane surface conditions
• Weather & visibility condition
10s
0s
10s
0s
9RF9MF
9LF
27RT27LT
18
LF
18
MF
18
RF
0L
T 0R
T
9RT9MT
9LT
27RF
27LF27MF
5RT5M
T5LT 31LF
31RFLDM Content:Example:Intersection
17General Assembly2007-01-30, Orbassano (CRF)
Summary of Main SP1 Achievements in 1st Year
• 3 Deliverables completed in due time:
- D1.2.1 Vehicle probe use cases- D1.2.2 System analysis of useful in-vehicle data- D1.2.3 Requirements for the SAFEPROBE platform
• Architecture designed jointly with other SPs- Client-server with brokers- Local Dynamic Map database, T/Q-APIs, query & notification,… (SP3)
• Local Dynamic Map Contents proposed
- Objects of static and dynamic layers- Trajectory prediction hypotheses- Situations/Events: timely & spatial relations betw. objects, trajectories,…
• Open Java Micro Traffic Simulator developed & published to consortium