Challenges for automated driving

Report
The Dutch Automated Vehicle Initiative:
Challenges for automated driving
Dr. R.(Raymond) G. Hoogendoorn
Assistant Professor Delft University of Technology
Challenge the future
1
1.
Automated driving:
definitions and impacts
The Dutch Automated Vehicle Initiative
2
The Dutch Automated Vehicle
Initiative (DAVI)
• What is the objective?
• To investigate, improve and
demonstrate automated
driving on public roads
The Dutch Automated Vehicle Initiative
Challenge the future
3
What is automated driving?
• Driver assistance:
• The driver permanently maintains either longitudinal or lateral
control. The other task can be automated or advisory.
• Partial automation:
• The system takes over longitudinal and lateral control. The
driver must still permanently monitor the system and be
prepared to take over control.
• High automation:
• The system takes over longitudinal and lateral control. The
driver is not required to permanently monitor the system, but
must be prepared to resume manual control.
• Full automation:
• The system takes over longitudinal and lateral control. In case
a take-over request is not carried out, the system will return to
a minimal risk condition by itself.
The Dutch Automated Vehicle Initiative
Challenge the future
4
What are the possible benefits of
automated driving?
Solve traffic jams
by increased
outflow
Reduction in
congestion
Increase traffic
stability
Improved travel
experience
Towards an
optimal distribution
over the network
Increase in traffic
safety (also
VRU’s)
Improved energy
efficiency
The Dutch Automated Vehicle Initiative
Challenge the future
5
2.
The Challenges
The Dutch Automated Vehicle Initiative
6
Challenges of automated driving
• Four main challenges:
• Technology;
• Human Factors;
• Traffic management applications;
• Traffic safety and legal;
The Dutch Automated Vehicle Initiative
Challenge the future
7
2a.
Technology
The Dutch Automated Vehicle Initiative
8
Technology challenges
• Environmental perception & communication;
• Automated vehicles are required to be
sufficiently aware of the driving environment as
well as the interaction with other road users;
• Sensing technologies are already capable of
identifying other road users, obstacles,
navigation paths, road signs and delineation
(Broggi & Zani, 2011);
• However, the reliability of sensing has to be
quantified;
• Furthermore the reliability has to be improved
for real life conditions (e.g. adverse weather
conditions and complex traffic scenarios);
The Dutch Automated Vehicle Initiative
Challenge the future
9
Technology challenges2
• Automation control strategies;
• Current control strategies mainly focus on
longitudinal control (e.g., ACC);
• Up to now, lateral control systems are
predominantly advisory (Visvikis et al., 2008);
• Automated control strategies have to be
developed for:
• Merging;
• Lane changing;
• Overtaking;
• Testing of these control strategies in real life
conditions;
The Dutch Automated Vehicle Initiative
Challenge the future
10
Technology challenges3
The Dutch Automated Vehicle Initiative
Challenge the future
11
2b.
Human factors
The Dutch Automated Vehicle Initiative 12
Human factors challenges
• In automated driving the human driver
becomes a supervisor of a system instead of
a manual controller of the vehicle;
• In partial and high automation, a capable
driver is still required to resume manual
control;
• Profound insight is needed into the quality of
the determinants of the quality of the
interaction of the driver with the automated
vehicle;
• Examples of the determinants are vigilance,
driver workload and situation awareness
(Brookhuis et al., 1991; Endsley, 1995);
The Dutch Automated Vehicle Initiative
Challenge the future
13
Human factors challenges2
• Currently little knowledge is available on these
determinants in case of driving in an
automated vehicle under real life traffic
conditions;
• Most knowledge in relation to behavior is
largely based on driving simulator studies;
• Furthermore, traffic safety may be affected by
adaptation effects in behavior of the
surrounding manually driven vehicles
(mixed traffic);
• It is however not yet clear to what extent
driving behavior of these vehicles is affected;
The Dutch Automated Vehicle Initiative
Challenge the future
14
2c.
Traffic management
The Dutch Automated Vehicle Initiative 15
Traffic management challenges
• The advent of V2I and V2V in combination
with automated vehicles offers a vast range of
possibilities to increase traffic flow efficiency;
• E.g., platooning and lane specific control;
• Rao and Varaiya (1994): framework for
controlling the movement of platoons;
• Steady state flow of over 6000 veh/hr;
• Knoop et al. (2012): lanes are generally
underutilized;
• Furthermore, currently traffic management
measures often result in suboptimal
performance since drivers disregard advice
or comply imprecisely (Risto & Martens,
2011);
The Dutch Automated Vehicle Initiative
Challenge the future
16
Traffic management challenges2
• A complication is however that until now
studies on automated traffic management
have predominantly been investigated using
microscopic simulation studies;
• E.g., Wang et al., 2012; Schakel et al., 2010;
Park et al., 2011;
• Empirical verification of the assumptions
underlying these theories and models on the
behavior of automated vehicles and drivers in
dynamic manoeuvers is urgently needed;
The Dutch Automated Vehicle Initiative
Challenge the future
17
2d.
Traffic safety and legal
challenges
The Dutch Automated Vehicle Initiative 18
Traffic safety and legal challenges
• The reliability of the sensors and human
aspects can be used to quantify benefits in
terms of traffic safety;
• However, an adequate quantification of traffic
safety does not yet exist;
• Benefits in terms of traffic safety is envisaged
to be explored using microscopic simulation
studies;
• However, currently models are predominantly
accident free. New models are needed!
The Dutch Automated Vehicle Initiative
Challenge the future
19
Traffic safety and legal challenges2
• But also many legal challenges are ahead;
• The vehicle should be offered to the Dutch
vehicle approval authority for admittance on
public roads;
• Finally substantial challenges exist in terms of
liability…
• Who is legally responsible when a crash
involving an automated vehicle occurs?
The Dutch Automated Vehicle Initiative
Challenge the future
20
3.
A closer look at DAVI
The Dutch Automated Vehicle Initiative 21
The Dutch Automated Vehicle
Initiative
• Delft University of Technology;
• Transport Institute;
• Robotics Institute;
• Connekt:
• Dutch Transport and Automotive sector;
• RDW:
• Legalization of automated driving;
• TNO and Toyota Motor Europe;
• Many Dutch and selected international partners;
• But we are always looking for new partners!!!
The Dutch Automated Vehicle Initiative
Challenge the future
22
The Dutch Automated Vehicle
Initiative2
• Partners investigating automation
Since 1990ties
• Inception DAVI
March 2013
• DAVI grant proposals
June 2013
• First full time DAVI researcher
August 2013
• Official Launch Innovatie-estafette
November 2013
The Dutch Automated Vehicle Initiative
Challenge the future
23
The Dutch Automated Vehicle
Initiative2
• Partners investigating automation
Since 1990ties
• Inception DAVI
March 2013
• DAVI grant proposals
June 2013
• First full time DAVI researcher
August 2013
• Official Launch Innovatie-estafette
November 2013
The Dutch Automated Vehicle Initiative
Challenge the future
24
4.
Thank you for your
attention!
The Dutch Automated Vehicle Initiative 25

similar documents