Autonomy Levels, Software & Artificial Intelligence Medium

Autonomy Levels & Software & AI (Medium)

1 / 16

Why do the data connections of autonomous cars need to be fast?

To stream music live

To share real-time sensor data

To receive phone calls

To contact a human in the case of an emergency

Fact:
AI-powered systems can reconstruct hidden areas and predict traffic movements to help the driver perceive their surroundings more accurately.

2 / 16

In the past, if your car had a software bug, you went to the mechanic. How do modern Software-Defined Vehicles fix this kind of bug?

Thanks to most modern engineering practices, there are no bugs that require a fix.

Over-the-Air updates, like your smartphone.

The car's AI identifies and fixes the problems automatically.

A automatically invoked hard reset of all components, restores a bugless state.

Fact:
Future C-V2X communication systems will use weak AI systems in many places to continuously optimize communication and provide excellent latency, reliability and throughput.

3 / 16

What is the main benefit of using Roadside Units (RSUs) to assist autonomous vehicles?

To provide faster internet for passengers

To extend the vehicle's line of sight and enhance autonomy

To charge the vehicle's battery wirelessly

To replace the vehicle's internal computing system

Fact:
Roadside unit sensors can detect objects behind large obstacles that a vehicle's onboard sensors would not be able to.

4 / 16

Which AI technique is most commonly used for object detection in AVs?

Decision trees

Convolutional Neural Networks (CNNs)

Genetic algorithms

Bayesian inference

Fact:
Autonomous vehicles use smart computer systems that learn from millions of driving experiences. This helps them notice things like pedestrians or other cars and make quick, safe decisions — often faster than a human driver.

5 / 16

What is the main goal of ODD-aware decision-making?

Predict energy use

Adapt planning to ODD limits

Optimize map compression

Switch modes based on latency

Fact:
AI systems detect when conditions shift outside the ODD and adjust decisions accordingly.

6 / 16

What makes autonomous software so complex to design

The math behind it is too complex

There exist no good programmers that can be instructed for it

It shall run on very small computers

Fact:
AI is capable of correctly analyzing and categorizing high quantities of data, such as the ones coming from the vehicle sensors, with great effectivenes when compared to other techniques

7 / 16

What is an AI model like "U-Net" typically used for in the context of autonomous driving, as described in Dem3.2?

Route planning

Semantic Segmentation (pixel-level classification)

Voice commands

GPS signal interference detection

Fact:
Software in an autonomous car must manage "mixed-criticality". This means ensuring that vital safety functions (like braking) always have priority and can never be disrupted by comfort functions (like the infotainment system).

8 / 16

What is the purpose of tracking in a vehicle management system?

Maintaining information about the identity, location, or state of vehicles over time

Changing the vehicle brand

Turning headlights into sensors

Making road signs digital only

Fact:
Vehicle management systems rely on timely information about vehicle position, status, and availability to support safe coordination.

9 / 16

How can location awareness help make digital vehicle access safer?

By reducing the chance that a distant signal is accepted as nearby

By removing the need for encryption

By making the vehicle invisible

By disabling all sensors

Fact:
Connected vehicle services combine software authorization, positioning, cybersecurity, and user experience design.

10 / 16

Why are compact sensors valuable for future vehicles?

They can be easier to integrate into vehicle bodies and electronic architectures

They remove the need for software

They make maps unnecessary

They reduce road signs

Fact:
Object detection and scene understanding turn raw sensor data into information that automated-driving software can use for decisions.

11 / 16

How does self-supervised learning benefit virtual sensing systems?

It eliminates the need for vehicle sensors entirely

It allows models to learn useful representations without requiring labeled data

It removes the need for hardware testing

It ensures that all datasets are perfectly balanced

Fact:
AI enables vehicles to interpret complex sensor data and infer missing or unmeasurable variables.

12 / 16

In SAE Level 3–4 autonomous vehicles, how do AI algorithms and software systems mainly contribute to vehicle operation?

They only process raw sensor data and do not affect control decisions

They enable real-time perception, decision-making, and navigation while supporting safe handover between human and automation

They are used exclusively for infotainment and user interfaces

They remove the need for redundancy and safety monitoring

Fact:
At SAE Levels 1–2, AI-based perception helps the driver by recognizing lanes, vehicles, and pedestrians, but the human remains fully responsible for driving decisions.

13 / 16

How does AI software typically identify a 'micro-sleep' or drowsiness using steering data?

By detecting constant, smooth steering

By identifying the lack of micro-corrections or sudden jerky movements

By measuring the speed of the vehicle

By checking if the turn signals are used

Fact:
Specialized software algorithms analyze biometric data throughout the day to evaluate a rider's energy level and detect signs of drowsiness or tiredness. These AI-enabled systems facilitate communication between the monitoring device and the vehicle's motor controller to execute preventive actions when needed. While cars are increasingly equipped with systems to detect distraction, the two-wheeled vehicles cannot be automated but can warn the driver in time to prevent accidents.

14 / 16

What is the main purpose of an occupancy grid map in autonomous vehicle perception?

To represent the environment as occupied or free space for safe navigation

To store high-definition map features like lane markings and traffic signs

To plan the route between cities

To record past sensor data for training AI models

Fact:
Spiking Neural Networks (SNNs) are emerging AI technique for event-based vision because they process data as discrete temporal spikes.

15 / 16

At which SAE automation level does the vehicle perform obstacle avoidance without continuous driver supervision?

SAE Level 1

SAE Level 2

SAE Level 3

SAE Level 4

Fact:
Spiking Neural Networks (SNNs) are emerging AI technique for event-based vision because they process data as discrete temporal spikes.

16 / 16

Why is accurate sensor alignment important for object detection?

It makes the dashboard brighter

It helps software estimate where objects really are around the vehicle

It improves the radio station signal

It reduces the vehicle weight

Fact:
Sensor fusion helps automated-driving software combine complementary information so that one sensor’s weakness can be balanced by another sensor’s strength.

The average score is 0%

Your score is

Autonomous Vehicles Quiz Game

Autonomy levels & software & AI – Medium

FOLLOW US