Safe Interaction, Connectivity and State
Hui Zhong Advanced Technology & Research, Volvo Group Trucks Technology
Objectives
Main goal is to safely deliver connected lifestyle functionality while also improving driver attention Overcome challenges related to driver distraction guidelines and regulation Meet increased consumer demand for connected lifestyle Integrate state-of-the-art on-board driver inattention monitoring
Volvo Group Trucks Technology Hui Zhong, SICS project, Resultatkonferens TSAF, September 21, 2016
Outline
SP1 HMI Guidelines and Validation Processes
SP2 Safe HMI designs SP3 Safe Connectivity SP4 Driver Inattention Monitoring
Volvo Group Trucks Technology Hui Zhong, SICS project, Resultatkonferens TSAF, September 21, 2016
SP1 and SP2
Research questions:
– What is the reliability and validity of performance criteria? – What is the feasibility of testing procedures? – What are the practical implications of performance testing on HMI design?
Volvo Group Trucks Technology Hui Zhong, SICS project, Resultatkonferens TSAF, September 21, 2016
A typical simulator experiment
Volvo Group Trucks Technology Hui Zhong, SICS project, Resultatkonferens TSAF, September 21, 2016
NHTSA test carried out at Volvo Cars
A full-scale NHTSA test with two test samples, in total 48 participants
Volvo Group Trucks Technology Hui Zhong, SICS project, Resultatkonferens TSAF, September 21, 2016
Acceptance criteria
No more than 3 out of 24 of the participants shall fail to meet each of the following criteria: Acceptance criteria 1 (AC1): For at least 21 of the 24 test participants, no more than 15 percent (rounded up to the next whole number) of each participant’s total number of eye glances away from the forward road scene have durations of greater than 2.0 seconds while performing the testable task one time
Acceptance criteria 2 (AC2): For at least 21 of the 24 test participants, the mean duration of each participant’s eye glances away from the forward road scene is less than or equal to 2.0 seconds while performing the testable task one time Acceptance criteria 3 (AC3): For at least 21 of the 24 test participants, the sum of the durations of each individual participant’s eye glances away from the forward road scene is less than or equal to 12.0 seconds while performing the testable task one time Volvo Group Trucks Technology Hui Zhong, SICS project, Resultatkonferens TSAF, September 21, 2016
What is the reliability and validity of performance criteria? Do two tests yield the same results? When comparing the test outcomes between two groups, it was found that for 6 of the 10 tested tasks, pass/fail outcomes differed between the two groups Are the criteria well founded?
Individual variations in glance strategy exist and that these variations seem to have a non-negligible influence on the NHTSA acceptance test criteria Are there better ways to calculate pass/fail criteria? Group means rather than the proposed frequency count of individual passes and failures, would effect between-group robustness in task compliance testing
Volvo Group Trucks Technology Hui Zhong, SICS project, Resultatkonferens TSAF, September 21, 2016
Experiments carried out at Volvo Trucks
1.2) Embedded occlusion vs glasses occlusion part 2 (difficult tasks) 1.3) Effects of simulator view (truck vs car) and display position (low/high) on glance measures 1.4) Occlusion vs eyetracking
Validation methods
1.1) Embedded occlusion vs glasses occlusion part 1 (normal tasks)
2.0) On road study with eyetracking, current FH truck => Expert evaluation of problematic tasks
2.2) Using sound to reduce distraction part 2: speech feedback for handwriting text input
2.3) Text presentation part 1: Effects of font type 2.4) Text presentation part 2: Effects of polarity (Black text on white bg vs. White text on black bg) 2.5) Text input method: Keyboard vs. Handwriting
Safe HMI recommendations, input to standards, guidelines Volvo Group Trucks Technology Hui Zhong, SICS project, Resultatkonferens TSAF, September 21, 2016
Safe HMI designs
2.1) Using sound to reduce distraction part 1: speech/sound feedback for menu navigation
Distraction-reducing auditory display
Can distraction from in-vehicle interfaces be reduced by adding guiding sound to e.g. menu browsing? Two sound concepts based on previous research developed: Earcons (musical sounds), Spearcons (speech-based sounds)
Earcons
Volvo Group Trucks Technology Hui Zhong, SICS project, Resultatkonferens TSAF, September 21, 2016
Spearcons
Total glance time 9.000
8.032
7.557
Earcon
Baseline
8.000 7.000 6.000 5.000
4.147 4.000 3.000 2.000 1.000 .000 Spearcon
Significant difference between Spearcons and Earcons/Baseline (p < .01)
Volvo Group Trucks Technology Hui Zhong, SICS project, Resultatkonferens TSAF, September 21, 2016
SP1-2 conclusions
Overall the work has
played an important role in the correspondence with HMI guidelines, e.g. ESoP, JAMA, ISO, NHTSA evolved design guidelines used within the product development process provided insights on the feasibility of test methods provided practical recommendations on safe HMI design
Volvo Group Trucks Technology Hui Zhong, SICS project, Resultatkonferens TSAF, September 21, 2016
SP3 Safe Connectivity No or limited “Driver Distraction” control over use of Apps directly on Smartphone (Use, HMI design, content) OEM Certified Apps
Head-Unit HMI
High control of App (Use, Content, HMI design, Notification, Driver distraction)
Volvo Group Trucks Technology Hui Zhong, SICS project, Resultatkonferens TSAF, September 21, 2016
Safe Connectivity Platform Safe HMI design recommendation
Assessment Tools
App approval process Safety API
Volvo Group Trucks Technology Hui Zhong, SICS project, Resultatkonferens TSAF, September 21, 2016
App contest Chalmers course: DAT255, Software Engineering Project, 7.5 HEC, HT2014 8-week competition (September to October 2014) 172 students
19 teams Challenge: Design, develop & assess Apps for truck drivers running on the Automotive-Grade Platform, & using the SICS project background & inputs
Volvo Group Trucks Technology Hui Zhong, SICS project, Resultatkonferens TSAF, September 21, 2016
SP4 Driver Inattention Monitoring
Investigate state of art Driver State Monitoring Camera System which
– monitors the driver’s head pose, gaze direction and eyelid closure – detects driver attention such as eyes on/off road, driver lap, instrument cluster, right/left side, micro sleep and asleep.
Volvo Group Trucks Technology Hui Zhong, SICS project, SAFER Thursday Seminar, January 28, 2016
Sensor Performance Testing
Test setup
Analysis
– 13 sticker test
– Uptime
– Re-initialization
– Re-initialization
– Naturalistic driving
– Noise
– Radio task
– Accuracy
– Blinking task
– On/Off road
Volvo Group Trucks Technology Hui Zhong, SICS project, SAFER Thursday Seminar, January 28, 2016
Test Result Summary
The overall system shows good reliability and accuracy. Critical to have a good mounting position Gaze Uptime is very low which makes it unusable in practice Head Uptime is pretty good, however not sufficient for more than basic on/off road detection
Volvo Group Trucks Technology Hui Zhong, SICS project, SAFER Thursday Seminar, January 28, 2016
Gaze Zone Estimation
Sensor data set:
– Head pose – Head pose + eye closure – Head pose + gaze data Calculate the probabilities of the driver’s visual focus at each of a set of defined sectors using Bayesian filtering
Gaze mapping from head rotation and eye closure using Gaussian processes shows improved gaze zone estimation performance compared to using filtered data Volvo Group Trucks Technology Hui Zhong, SICS project, SAFER Thursday Seminar, January 28, 2016
Thanks for your attention! Hui Zhong Advanced Technology & Research Volvo Group Trucks Technology
Volvo Group Trucks Technology Hui Zhong, SICS project, Resultatkonferens TSAF, September 21, 2016