Page 24 - AI Ready – Analysis Towards a Standardized Readiness Framework

AI Ready – Analysis Towards a Standardized Readiness Framework

further study” covers the analysis of use case scenarios. We cluster the use cases based on the
domain they involve.

4.2 Traffic Safety

In this section, several use cases related to traffic safety are studied. AI technologies are
integrated into autonomous and remote-driven vehicles with the aim to improve both drivers
and pedestrians’ safety on the road.

4�2�1 Platooning

This use case [24] [23] involves autonomous or remote-driven vehicles such as enterprise
vehicles, in-campus vehicles, carts, and mover trucks in stores, factories, or maritime ports.
To control the optimal distance vehicles, proximity and distance sensors using ultrasound are
applied. Due to the strict requirements of latency and bandwidth, a 5G network is needed.
Private 5G such as campus networks managed and operated by the enterprise, may be used.
Bluetooth might be used for communications between each unit. To facilitate communication,
standards such as Dedicated Short-Range Communication (DSRC) [47] and IEEE 802.11p [42]
are considered. Human supervision is still needed to minimize the risks in deployment. In
addition, roaming cases across borders should be planned in advance [29].

4�2�2 Driving under Intoxication (DUI) Detection

Driving under intoxication [8] [10] has different standards in different countries, so it is essential
to obtain the test parameters and thresholds such as blood alcohol level accordingly for
standard field tests. In terms of in-vehicle measurement, various inertia, response times, and
signal processing on the controls should be considered. Based on the chest movement, medical
data such as heartbeat, breathing, and even blood alcohol level might be inferred.

4�2�3 Autonomous Emergency Braking-based (AEB) Collision Avoidance

This use case used [7] algorithms to predict and avoid collisions by collecting data from a sensor
fusion and multi-modal for commercial vehicles. Specifically, blind spots, lane merging, and
direction of movement are noted to complement the angle detection data. In case of collision,
safety accessories such as seat belts and airbags are to protect the drivers and passengers. To
test the AEB system [32][33], an international standard issued by the Society of Automotive
Engineers (SAE) is referred to. In addition, driver preferences and parameters for the level of
alert should be considered according to the different driving habits of drivers.

4�2�4 Pedestrian Safety

Smart vehicles, sensors such as Lidar and cameras, networks with low latency and high
throughput, AI/ML feedback loops, and roadside units (RSUs) like smart traffic signals play an
important role in this use case. The level of impact of the use case, its potential influence, and
predictive human intervention (preventive and post-facto interventions) are studied. Decision-
making aspects, such as sensor deployment responsibilities, subsidies, and goal setting for
deployment timelines, are critical.

17