> Markdown version of [/videos/628-automated-driving-why-is-it-so-hard-to-introduce](https://www.wearedevelopers.com/videos/628-automated-driving-why-is-it-so-hard-to-introduce). Every page supports `.md` or `Accept: text/markdown`. Links point to the HTML versions so they work for humans too. Agent guide: [/agents.md](https://www.wearedevelopers.com/agents.md).

---

# Automated Driving - Why is it so hard to introduce

When a decelerating car pitches downward, sensors can scan too low and detect ghost objects. Learn why overcoming closed-loop perception challenges requires radically restructuring legacy automotive engineering teams.

- **Speakers:** Sayed Bouzouraa
- **Event:** World Congress 2023
- **Published:** August 11, 2023
- **Duration:** 31:23
- **URL:** https://www.wearedevelopers.com/videos/628-automated-driving-why-is-it-so-hard-to-introduce

## Summary

The transition from assisted driving to fully automated driving presents massive engineering and organizational challenges that cannot be solved by simply iterating on legacy systems. While Level 2 assisted driving is reaching market saturation, pushing it to an optimized "Level 2++" is an endless trap. True automated driving (Level 3 and beyond) fundamentally shifts liability to the carmaker, creating an "automation dilemma" where increased automation capability decreases a human driver's ability to quickly regain control. To release a commercially viable product, autonomous software engineers must prove a positive risk balance against an incredibly competitive safety baseline, such as the ultra-low fatality rates of human drivers on highways.

At the core of this challenge is the sheer complexity of environment perception and the difficulty of mapping a long tail of unpredictable corner cases. Evaluating sensors in an open-loop vacuum fails to account for closed-loop sensor behavior, where the vehicle's reaction dynamically alters its own perception. For example, if a car slightly decelerates to investigate a false reading, the vehicle's chassis pitch angle lowers; this forces sensors to scan closer to the ground, which can incorrectly confirm non-existent "ghost" objects. Rather than wasting compute resources trying to perfectly estimate the entire world or simulate endless sensor effects, autonomous development must pivot toward robust anomaly detection and establishing open industry marketplaces for safety-critical edge cases.

To overcome these technical hurdles, a radical shift in traditional automotive organizational structures is required. Because autonomous software touches nearly every mechanical and digital subsystem of a vehicle, continuing with isolated component silos inevitably destroys system cohesion. Effectively managing these dependencies requires deploying cross-functional "event chain teams" that take end-to-end agile responsibility—from initial sensor input to final mechatronic actuator. By tightly coupling software engineering, functional safety, and systems engineering, legacy automakers can successfully build resilient, hands-free automation architectures.

**Keywords:** automated driving challenges, level 3 automation, advanced driver-assistance systems, automation dilemma, corner case rate prediction, sensor perception anomalies, lidar ghost objects, closed-loop sensor behavior, automotive systems engineering, cross-functional agile teams, positive risk balance kpis, event chain teams, functional safety standards, autonomous vehicle development

## Chapters

1. **The reality gap between self-driving concepts and availability** (00:02) — A personal story highlights the discrepancy between marketing concepts and actually available autonomous vehicles.
1. **Audi progressive milestones in automated driving and perception** (05:33) — Previous historical autonomous projects reveal the evolution and complexities of shifting driving responsibility to the vehicle.
1. **Transitioning product focus from assisted to automated driving** (07:09) — Assisted driving is reaching performance saturation while automated functionality introduces disruptive challenges regarding responsibility shifts.
1. **Proving positive risk balance against human highway performance** (12:12) — Releasing an autonomous system requires proving safety and legal compliance that surpasses human highway driving statistics.
1. **Discovering and sharing edge cases for automated driving** (14:55) — A proposed marketplace for sharing safety-critical edge cases helps models tackle complex environmental variables.
1. **Interpreting raw sensor data and targeting anomaly detection** (19:26) — Lidar and camera vulnerabilities emphasize the need for targeted anomaly detection rather than fully estimating open-loop environments.
1. **Managing closed-loop system reactions to ghost object artifacts** (24:55) — Reacting to false environmental artifacts changes vehicle pitch angles and creates self-confirming radar detection loops.
1. **Adapting organizational structures for autonomous systems engineering teams** (26:54) — Cross-functional event-chain teams replace traditional departmental silos to better manage complex dependencies across vehicle hardware components.
1. **Addressing ethical dilemmas and automotive technology market competitiveness** (29:59) — Collecting specific scenario data addresses ethical edge cases while established safety engineering provides a regional competitive advantage.

## Related Moments

- [Navigating the levels of vehicle autonomy](https://www.wearedevelopers.com/videos/22-how-to-develop-an-autonomous-car-end-to-end-robotic-drive-and-the-mobility-revolution) (from "How to develop an autonomous car end-to-end: Robotic Drive and the mobility revolution")
- [Advancing autonomous driving capabilities with specialized software talent](https://www.wearedevelopers.com/videos/1360-inside-mercedes-benz-how-cio-katrin-lehmann-is-empowering-5-000-developers-and-driving-digital-change) (from "Inside Mercedes-Benz: How CIO Katrin Lehmann is Empowering 5,000 Developers and Driving Digital Change")
- [Audience questions on self driving evolution and legal progress](https://www.wearedevelopers.com/videos/258-on-developing-smartphones-on-wheels) (from "On developing smartphones on wheels")
- [Introduction to the speakers and topic](https://www.wearedevelopers.com/videos/22-how-to-develop-an-autonomous-car-end-to-end-robotic-drive-and-the-mobility-revolution) (from "How to develop an autonomous car end-to-end: Robotic Drive and the mobility revolution")
- [Audience Q&A on autonomous driving models and data](https://www.wearedevelopers.com/videos/519-finding-the-unknown-unknowns-intelligent-data-collection-for-autonomous-driving-development) (from "Finding the unknown unknowns: intelligent data collection for autonomous driving development")
- [Differences between levels of automated driving responsibilities](https://www.wearedevelopers.com/videos/1082-code-to-road-in-12-hours) (from "Code to Road in < 12 hours")

## Related Articles

- [How software is steering vehicle technology](https://www.wearedevelopers.com/magazine/515-how-software-is-steering-vehicle-technology)
- [Exploring AI: Opportunities and Risks for Developers](https://www.wearedevelopers.com/magazine/522-exploring-ai-opportunities-and-risks-for-developers)
- [Stephan Gillich - Bringing AI Everywhere](https://www.wearedevelopers.com/magazine/489-stephan-gillich-bringing-ai-everywhere)
- [Transforming Software Development: The Role of AI and Developer Tools](https://www.wearedevelopers.com/magazine/527-transforming-software-development-the-role-of-ai-and-developer-tools)

## Related Jobs

- [Staff Software Engineer, Copilot Experiences](https://www.wearedevelopers.com/jobs/ext/164361-staff-software-engineer-copilot-experiences) at **GitHub**
- [Senior AI Agent Software Engineer (Go, Python) (m/f/x)](https://www.wearedevelopers.com/jobs/48277-senior-ai-agent-software-engineer-go-python-m-f-x) at **Dynatrace**
- [AI Operations Manager (all genders)](https://www.wearedevelopers.com/jobs/48263-ai-operations-manager-all-genders) at **envelio**
- [Principal Software Engineer, Enterprise AI Platform](https://www.wearedevelopers.com/jobs/ext/1467292-principal-software-engineer-enterprise-ai-platform) at **GitHub**
- [Staff Developer Advocate, GitHub Security Lab](https://www.wearedevelopers.com/jobs/ext/1921051-staff-developer-advocate-github-security-lab) at **GitHub**
- [AI Software Engineer (Germany)](https://www.wearedevelopers.com/jobs/48317-ai-software-engineer-germany) at **Sunhat**