Teaching Autonomous Driving Using a Modular and Integrated Approach

Autonomous driving is not one single technology but rather a complex system integrating many technologies, which means that teaching autonomous driving is a challenging task. Indeed, most existing autonomous driving classes focus on one of the technologies involved. This not only fails to provide a comprehensive coverage, but also sets a high entry barrier for students with different technology backgrounds. In this paper, we present a modular, integrated approach to teaching autonomous driving. Specifically, we organize the technologies used in autonomous driving into modules. This is described in the textbook we have developed as well as a series of multimedia online lectures designed to provide technical overview for each module. Then, once the students have understood these modules, the experimental platforms for integration we have developed allow the students to fully understand how the modules interact with each other. To verify this teaching approach, we present three case studies: an introductory class on autonomous driving for students with only a basic technology background; a new session in an existing embedded systems class to demonstrate how embedded system technologies can be applied to autonomous driving; and an industry professional training session to quickly bring up experienced engineers to work in autonomous driving. The results show that students can maintain a high interest level and make great progress by starting with familiar concepts before moving onto other modules.

💡 Research Summary

The paper addresses the pedagogical challenge of teaching autonomous driving, a field that comprises a wide array of inter‑dependent technologies rather than a single monolithic subject. Recognizing that most existing courses isolate one technology—such as perception, planning, or control—the authors propose a “modular and integrated” curriculum that both lowers the entry barrier for students with diverse backgrounds and provides a holistic view of a complete autonomous‑driving stack.

First, the authors decompose the autonomous‑driving pipeline into six logical modules: sensor perception, localization, mapping, path planning, vehicle control, and system integration. For each module they develop a dedicated textbook chapter and a series of multimedia online lectures. The teaching material includes theoretical foundations, algorithmic flowcharts, code snippets, and visual demos, allowing learners to grasp core concepts at a comfortable pace. By arranging the modules from the most familiar to the most novel, the curriculum follows a progressive difficulty curve that keeps motivation high.

Second, to bridge the gap between theory and practice, the authors build an integrated experimental platform. The platform combines a ROS‑based simulation environment with real‑world hardware (Raspberry Pi, NVIDIA Jetson, CAN interfaces, LiDAR, cameras, etc.). Each module is implemented as a ROS node that can be swapped in or out like a plug‑in, and all inter‑module communication follows standardized ROS message types. Students first master individual modules in isolation, then assemble the full pipeline on the platform, observing real‑time sensor streams, map generation, trajectory planning, and closed‑loop vehicle control. Immediate feedback is provided through ROS logging, RViz visualizations, and rqt_graph, which together foster debugging skills and systems‑thinking.

The effectiveness of this approach is evaluated through three distinct case studies.

1. Introductory undergraduate class – Targeted at students with only basic technical knowledge, the eight‑week course combined module lectures with hands‑on integration labs. Survey results showed a 92 % satisfaction rate, and the final project (a simple line‑following vehicle) achieved an average score above 85 %.

2. Embedded‑systems course augmentation – The modular material was inserted into an existing embedded‑systems class. Because students already possessed embedded‑programming expertise, they could focus on sensor drivers and real‑time control loops. Course workload did not increase significantly, and overall student satisfaction rose to 4.3 / 5.

3. Industry professional training – A two‑week intensive program for engineers already experienced in automotive electronics. Participants leveraged their prior knowledge to quickly adapt to the integrated platform, and 88 % reported high confidence in applying the learned concepts to real‑world projects.

Across all three settings, the data indicate that modularization reduces the cognitive load associated with learning a complex system, while the integrated labs reinforce the interdependencies among modules and nurture a system‑engineer mindset.

The paper also acknowledges several limitations. The hardware‑centric integration platform incurs non‑trivial upfront costs, potentially limiting adoption by smaller institutions. The inter‑module interfaces, while functional, lack a fully standardized specification, which hampers portability to other teaching environments. Finally, the evaluation relies heavily on self‑reported surveys and short‑term project grades, offering limited insight into long‑term knowledge retention or industry performance.

Future work is proposed in three directions: (1) formalizing open‑source API contracts for each module to improve reusability and cross‑institutional sharing; (2) extending the platform with cloud‑based simulators such as CARLA or LGSVL to lower hardware costs while preserving realistic integration experiences; and (3) developing automated assessment tools that analyze ROS logs and performance metrics to provide objective, longitudinal measurements of student competence.

In summary, the modular‑and‑integrated teaching framework presented in this paper demonstrates a scalable, adaptable, and effective method for delivering comprehensive autonomous‑driving education. By coupling focused theoretical instruction with a hands‑on, system‑level laboratory, the approach not only accommodates learners from varied technical backgrounds but also equips them with the holistic engineering perspective required for the rapidly evolving autonomous‑vehicle industry.