Towards the Self Healing Vehicle

1
Towards the Self Healing Vehicle

• 7th November 2006
• Presented by Mark Amor-Segan,
• (Ross McMurran, Peter Jones, Gunny Dhadyalla)

2
Towards the Self Healing Vehicle

• Introduction Why do we need a Self Healing
  Vehicle?
• What is a Self Healing Vehicle?
• Self Healing Vehicle - Requirements
• Future Research Opportunities
• Questions

3
Why do we need a Self Healing Vehicle?
Electronic Towbar
Adaptive Cruise Control
Function Growth

• The vast majority of new technology looks like
  this..

Communications Network
4
Why do we need a Self Healing Vehicle?

• 2 binary inputs means 4 test cases

Inp 1
Out 1
Inp 2

• 10 binary inputs means 1024 test cases

• Inputs represent a range of values not just 0 or
  1 (e.g. speed, temperature) then its infeasible
  to rely testing.

5
Why do we need a Self Healing Vehicle?

• Increasing complexity of embedded vehicle
  electronics and software systems.
• Electronics software account for over 40 of
  the total value of a premium vehicle increasing
  every year
• Cost of warranty
• Customer Satisfaction

6
Moores Law
7
500kbps
125kbps
26Mbps
8
Why do we need a Self Healing Vehicle?
9
Why do we need a Self Healing Vehicle?

• Conclusion
• There needs to be a step-change in fault
  management capabilities to keep pace with
  developments
• Work already underway at IARC in this area

10
What is a Self Healing Vehicle?

• A vehicle with the ability to autonomously
  predict or detect and diagnose failure
  conditions, confirm any given diagnosis, and
  perform appropriate corrective intervention(s),
  including the use of telematics to interact with
  external service providers and infrastructures.

11
Self Healing Vehicle Concept
Vehicle Data
Confirmed/Classified Failure Information
Diagnostic Prognostic Information, Data
Logging
Prognostic Diagnostic Monitor
Interrogation commands
Distributed Vehicle Electronics System
Intelligent Rectification Manager
Remote Telematics Support System
Corrective Action or Intervention
Intervention Initiator
Verification of Intervention outcomes
Failure Details Recommended Intervention
SW Downloads, Enhancements Upgrades to
diagnostic System, Remote commands.
In-vehicle Fault Management System
12
Self Healing Vehicle - Timescales?
13
Self Healing Vehicle Some general requirements
to consider

• Ability to work in real-time, within the vehicle
• Able to deal with whole vehicle electronic
  software systems
• Able to confirm any given diagnosis
• Intervention must be safe, reliable, effective
• Able to deal with new scenarios, uncertainty,
  missing or contradictory data
• Not an exhaustive list

14
Self Healing Vehicle - Current diagnostic
capabilities

• Limited On Board Diagnostics (OBD)
• Generally off-line, use historical stored data
  for post-processing
• External diagnostic tools, skilled engineers,
  labour intensive
• Deal with single fault single root cause
  (guided diagnostics, fault-tree).
• Not good where there is uncertainty, missing, or
  contradictory data

15
Self Healing Vehicle Current research work

• Model-based diagnostics
• Real-time monitoring diagnosis
• Bayesian Belief Network Diagnostics
• Multiple or distributed faults
• Able to deal with uncertainty
• Can suggest further observations to improve
  diagnostic outcome,
• Self learning
• Knowledge or Case-based diagnostics
• Scalable to large systems
• Able to reason about new or unknown failure
  conditions

16
Self Healing Vehicle Intervention Management

• Intervention Management
• Taking corrective action following confirmed
  diagnosis
• Cure can be worse than the disease
• Many examples of cascade faults through multiple
  dependencies in distributed systems
• Must understand and model system interactions
  dependencies

17
Self Healing Vehicle Intervention Management

• What are implications of criticality of fault on
  fault management strategy?
• Mission/Safety critical, High/Medium/Low Priority
  
• How does nature and type of fault influence the
  design of fault management strategies?
• Software fault/corruption, Controller hardware
  fault
• How would use of telematics influence design of
  fault management strategy
• Remote service centre support

18
Self Healing Vehicle Current intervention
capabilities

• Very little intervention exists today
• Safe modes (limp home mode)
• No real application of intelligent intervention
  for fault mitigation
• New area of research for automotive

19
Self Healing Vehicle Intervention Management

• What would a typical intervention strategy look
  like?
• Download new software images
• Movement initialisation of application on an
  alternative or standby controller
• Re-allocation of sensors, actuators to new
  controller
• Need suitable vehicle system and component
  architecture to enable this to happen

20
Self Healing Vehicle Concept
Vehicle Data
Confirmed/Classified Failure Information
Diagnostic Prognostic Information, Data
Logging
Diagnostic Monitor
Interrogation commands
Distributed Vehicle Electronics System
Intelligent Rectification Manager
Remote Telematics Support System
Corrective Action or Intervention
Intervention Initiator
Verification of Intervention outcomes
Failure Details Recommended Intervention
SW Downloads, Enhancements Upgrades to
diagnostic System, Remote commands.
In-vehicle Fault Management System
21
Self Healing Vehicle Architectural changes

• How do we design a cost effective, realisable
  vehicle architecture to remove direct linkages
  between controller, sensor and actuator?
• Sensors actuators can be associated with any
  controller
• How do we design a generic, adaptable, re-usable
  controller architecture?

22
Self Healing Vehicle Architectural changes
Sensor
Sensor
Actuator
Actuator
ECU
ECU
Sensor
Sensor
Actuator
Actuator
?
CAN Bus
CAN Bus
Sensor
Sensor
Actuator
Actuator
ECU
ECU
Sensor
Sensor
Actuator
Actuator
23
Self Healing Vehicle - Conclusions

• Current situation cannot continue
• Increasing complexity and criticality of
  applications
• Despite improvements in validation techniques,
  faults will still get to market,
• Electronics software will fail.
• Human-assisted monitoring, maintenance, and
  intervention will become prohibitively costly,
  unacceptably slow, and sometimes ineffective.
• An intelligent vehicle needs to play a more
  proactive role in fault management

24
Self Healing Vehicle - Conclusions

• Large problem space
• Plenty of opportunities
• UoW potential in a number of areas
• Huge potential benefits if Self Healing Vehicle
  can be realised
• Both for business and customer
• Whole-life vehicle costs
• Keeping the vehicle on the road

25
Potential areas of WIMRC Research
Advanced Prognostics/Diagnostics
Telematics
Intelligent Intervention Management
Whole-vehicle system modelling
26
The Self Healing Vehicle

• Thank you
• Any Questions?

27
Model-based diagnostics

• Successful Case Studies
• Cam Profile Switching (CPS)
• Engine Misfire Detection

28
DiagnosticsBayesian Guided Diagnostics
29
Knowledge or Case-based diagnostics