Jeehoon KI - National strategies for the development of autonomous vehicles. A comparative analysis of Korea, Japan, France

Submitted by Géry Deffontaines, GERPISA on Thu, 11/21/2019 - 01:00

Gerpisa monthly seminar

Publication Type:

Compte Rendu / Report

Authors:

Source:

Proceedings of the Gerpisa monthly seminar, Number 249, Paris - CCFA (2019)

Notes:

Jeehoon KI

Full Text:

Brief summary of the presentation

Jeehoon KI presented his comparative study of public policies in three countries to develop and deploy autonomous vehicles and/or autonomous driving technologies, based on an analysis of documents and framework texts.

It appears that the Korean plan is clearly directed towards industrial competitiveness and export opportunities. In Japan, the ambition to remain at the forefront of technology, which also motivates highly organized planning, is combined with the concern to use this form of robotization to meet the challenge of providing transport services to regions in demographic decline and an aging population. In France, the AV was first identified as a support product for reindustrialization before a national strategy was put in place to coordinate the various ministries. Social acceptability issues and shuttle experiments are given high priority.

The most significant slides and illustrations are included in these proceedings, but the presentation material can nevertheless be downloaded at the following link :http://gerpisa.org/system/files/Ki_-_2019.09.27_-_JG_ndeg249_-_slides.pdf

Glossary

VA / AV autonomous vehicle.
ADUS autonomous driving for universal service
C-ITS Cooperative Intelligent Transport Systems
MaaS mobility as a Service

Many countries are preparing for the AV (autonomous vehicle) era by addressing the various dimensions essential to its development: technologies, experiments, infrastructure, legislation and regulation. Public policies have a crucial role to play in this type of socio-economic transformation, especially in the event of the emergence of a new paradigm, by establishing the "rules of the game". What are the objectives of the policies implemented (economic, societal, environmental)? What is the best level of public intervention in terms of safety regulations, choices between competing technological options, especially for infrastructure?

The objective of the project carried out at the France-Japan Foundation (EHESS), the results of which are presented here, was to analyse in a comparative way the strategies and public policies relating to the AV in South Korea, Japan, and France.

The theoretical framework used is that of the "sectoral system of innovation" analysed by Franco Mallerba(According to which technological change takes different trajectories and speeds depending on the sector, in a game where actors and institutions play an essential role in innovation. The ISS analytical framework considers that the forms of innovation in the different sectors are affected by three sets of factors: technologies, networks of actors, institutions [additional note]). In the diagram proposed during the presentation, the interactions between public actors (administrations and universities) and private actors (firms), but also demand and markets, on the one hand, and the evolution of available technologies, on the other hand, are considered. Institutions (standards, practices, legislation) are the product of these games.

The methodology of the survey proposed is essentially based on the use of primary and secondary sources linked to grey literature : interviews with public policy makers, framework documents and other policy white papers, research reports, sectoral notes (e.g. note de France Stratégie, plan nouvelle France Industrielle, KPMG report, etc.), but also legal texts (e.g. draft LOM law)

If the project aims to compare the three countries, it is mainly the Korean case that is documented.Si le projet vise à comparer les trois pays, c’est surtout le cas coréen qui est documenté.

1. The Korean case

To give an idea of the topicality of the issue in Korea, Jeehoon Ki measures the "AV phenomenon" through a bibliometric study of press articles. There is clearly an intense and constantly growing topicality of the issue since 2015 (a threefold increase in the number of articles devoted to the subject in the news media).

Jeehoon KI introduces us to a group of players - only the largest of whom are known worldwide - who are working in Korea on the development of the autonomous vehicle. They come from various sectors - research and public authorities, technology or IT firms, manufacturers and equipment manufacturers. Some are well-known multinationals (Samsung, LG, KIA, Hyundai, Samsung, Ssangyong), others are start-ups. Among the public research centres, Katech, and ETRI (Electronics and Telecom Research Institute).

Source: Jeehoon Ki (2019)

By focusing on official plans and public policy framework documents, Jeehoon KI allows us to witness the rise of the VA strategy in South Korean economic planning:

In 2012, the "recognition" level, in a five-year automotive plan, which mentions the AV among R&D topics (but not a priority at the time), and an "IT convergence" strategy that mentions the AV as convergence between automotive and information technologies
From 2013, an increase to the "selection" phase in various scientific and technical, industrial development plans, which identify "smart cars" among 120 "national strategic technologies", "driverless transport" among the 15 "future IT services", "AVs" among the 13 technological development fields, "smart cars" among the 19 strategic fields of "future growth engines";
The transition to concrete planning takes place in 2015 with a detailed plan for R&D in areas related to "future growth engines" and action plans to support the commercialization of the AV
Finally, the acceleration phase in 2017-18 with the reaffirmation of the AV as a strategic area among the "innovation growth engines" and an update of the action plans within the framework of the "innovation growth strategy for AV and EV industries".

From 2015 to 2018, the share of public R&D expenditure allocated to AV-related projects increased from 0.1 to 1%. It should be noted that these plans seem to be mainly oriented towards applied research and industrial opportunities.

Three ministries are in charge of VA policies:
- The Ministry of Trade, Energy and Industry (MoTIE -Ministry of Trade, Industry, & Energy), which covers R&D expenditure in AV components (sensors...);
- The Ministry of Land, Infrastructure and Transport (MoLIT), in charge of safety standards, road infrastructure (including communication), the legislative and regulatory system governing the deployment of the AV;
- The Ministry of Science and Telecommunications (MSIT-Ministry of Science and IT), in charge of R&D investments in AI, information technology, data protection.

The policy roadmap details in three points the objectives of the Korean government:

- "level 3" vehicles on the roads in 2020 (now 2022);
- construction of infrastructure for "level 4+" VAs from 2022 onwards;
- circulation of "level 4+" VAsaround 2030 (mitigated by the speaker: "if the technologies are available")

The package of measures articulates 5 fields of action, here reported with their timetable:

1. R&D: technology and service development:

image 3 - Korea - MOLIT policy roadmap R&D

2. Infrastructure :
image 4 - Korea - MOLIT policy roadmap infrastructures

3. Technology demonstration:
image 5 - Korea - MOLIT policy roadmap démonstrateurs techno

4. Regulations and systems :
image 6 - Korea - MOLIT policy roadmap réglementations

5. Cultivation of Professionals :
image 7 - Korea - MOLIT policy roadmap formation

Calendar of AV Policy Actions (May. 2015, Feb. 2018), MoLIT

Source: JeehoonKi (2019) quoting MOLIT

As is the case in other countries, the Korean strategy is reflected in the development of test infrastructures, temporary permits to conduct operational tests, support for R&D, etc: :

- The Korean AV test site, called K-City, completed in December 2018, would be three times larger than M-city (U. of Michigan) to which its name refers. It would recreate most of the existing road environments - multi-lane highways with tolls, city centres, outskirts, car parks, but also, which is interesting for a country that does not have them and reflects a clear export ambition, roundabouts - would allow V2X interface tests to be carried out with 5G technologies, and would already be upgraded to allow tests on "level 5" vehicles (bad weather conditions for example).

Source: Jeehoon Ki (2019) quoting MOLIT

- Operational road tests began in February 2016 over 300 km, before the perimeter was extended, in November 2016, to all roads in the country, with the exception of a few protected areas for the most vulnerable pedestrians. At the end of 2018, 52 AVs were being tested on the roads, a third of which were implemented by Hyundai.

Source: JeehoonKi (2019) quoting MOLIT

- Finally, with regard to R&D support, the budget allocated to VA, identified as "next growth engine technology", has increased from €80 million in 2018 to €120 million in 2019. Efforts are focused on 9 critical technologies (low cost LiDAR, multiband radar, sensors, etc.). MIST R&D expenditures for 5G are also to be taken into account. The table below gives an idea of the main programmes carried out by the MoTIE, MSIT, MoLIT ministries

Source: Jeehoon Ki (2019)

Although it is not possible to say that the breakdown is the same for programmes related to the VA strategy, it is instructive to look at the types of organisations benefiting from public R&D expenditure in Korea: government research institutes (GRIs) and national laboratories account for almost 50%, universities for around 25%, and the balance (just under 25%) is allocated to private companies, two thirds of which for "small firms" (although it is not possible to know their size). According to Jeehoon KI, helping large, well-known firms outside the country would be relatively unpopular, so the government support they receive is reduced.

Beneficiaries of public funds for R&D.
Source: Lee based on OECD (2019)

Another interesting element is the final objective of public R&D expenditure, through a 6-pole analytical radar (support to universities, economic development, health/environment, education and society, aerospace, others): as Jeehoon KI has repeatedly pointed out, Korean public R&D expenditure is mainly oriented towards industrial benefits to support competitiveness and economic growth.

Beneficiaries of public funds for R&D.
Source: Lee based on OECD (2019)Source: Lee d’après OCDE (2019)

The adaptation of the law and regulations has also followed this path from the most general to the most detailed:
- The amendments to the motor vehicle management act made in 2015 defined the AV and created the legal basis for road tests;
- The Autonomous vehicle act, adopted in April 2019 with effect from 2020, classifies AV levels and lays the foundations for various public policies;

- A roadmap of regulatory changes to be implemented was also issued at the end of 2018: it aims to be proactive, in order to organise the development trajectory of the AV, and no longer simply to wait for requests from manufacturers on a case-by-case basis. It is sequenced over time to take into account a timetable corresponding to the deployment of successive levels of AV, and covers driver obligations, the objectives of on-board technologies, regulations applicable to accident cases, insurance, and finally the question of infrastructures (data collection, V2X standards...)

Source: Jeehoon Ki (2019)

Adaptation of infrastructures:

- The first step consists in the construction of 5G-compatible C-ITS (Cooperative Intelligent Transport Systems) demonstration infrastructures in Seoul by the end of 2020, followed by their extension to certain motorways.

- This will be followed by high-definition mapping, which should cover 5500km of the main network by 2020, then the entire road network (100,000km) by 2030. A high-precision GPS system must accompany this effort, by 2020 on motorways and in major cities, at national level by 2025.

The question of acceptability:

- First of all, it is interesting to note that there are only two trials in progress in Korea: since 2018, autonomous shuttles in the smart city Pangyo - which run at 25km/h on 5km - and three autonomous shuttles (including two Navya...) in Daegu at the end of 2019. The city of Pangyo also organised an AV motor show in 2017 and 2018, sponsored by the relevant ministries. However, the speaker insists on this point: priority is not given to demonstration experiments in Korea. In Seoul, a district inaugurated the first C-ITS test bench, based on 5G technology.

While these initiatives may seem anecdotal, they nevertheless give an idea of Korea's lead in 5G deployment and a design that gives priority to C-ITS technologies over a "societal" approach of tje AV. This is Jeehoon KI's overall conclusion: the first objective of the AV development in Korea is industrial, with the explicit goals of future strategic growth engineering to place Korea among the top three industries in the industry. There is less talk of using the AV to address social or environmental issues (congestion, ageing society, safety...). The focus is therefore on the AV as a product, rather than on the services it could help develop, as evidenced by the small number of service experiments - just one government-sponsored autonomous shuttle trial.

Jeehoon Ki will insist during the questions on the catching up mentality in economic policies in Korea, focused on exports, and on a rather dirigist apprehension of technological choices, which cannot be left to the market alone.

2. The Japanese case

Japan has been Korea's "oldest" country in terms of economic, industrial and technological development. The research results presented by Jeehoon KI allow us to see both commonalities in the thoroughness and ambition of planning, and elements of divergence in particular in the expected impacts of the development of the VA.

Let us first recall the different institutions involved in economic, industrial and R&D policy in Japan: various administrative and executive levels are involved in Japan in the definition and implementation of public policies in economic, industrial, R&D and infrastructure matters... of which the AV can be an objective, or which conditions its development:

- The well-known Minister of Economy, Trade, and Industry (METI);

- The Ministry of Land, Infrastructure, Transport, and Tourism (MLIT);

- And at both the governmental and administrative political levels, the cabinet office is responsible for the National Police Agency, administered by the National Public Safety Commission, and the Ministry of Internal Affairs and Communication

There may therefore be plans with a general political objective (economic recovery) and others that are part of a more long-term industrial policy and technological choices.

At the origin of VA development policies in Japan, there are various government and private initiatives from 2013 onwards:

- In 2013, the autopilot system council (part of MLIT) is developing a roadmap for the development of autonomous driving, with a focus on motorways
- In 2014, the Prime Minister's Office is launching an R&D programme focused on AV called "SIP-adus" (strategic innovation promotion programme - autonomous driving for universal service); while an "intelligent transport system" roadmap is launched
- In 2015, there is an explicit reference to AV in the national growth strategy of PM Shinzo Abe. A panel on business strategies in autonomous driving is established by METI and MLIT. The objective is twofold: to ensure Japanese competitiveness in the field of autonomous driving systems and to respond to various societal problems at the global level. An "Action Plan for Realizing Automated Driving" is published in 2017 and updated in 2018. The resulting roadmap provides a detailed timetable for passenger cars, freight logistics, and personal transport services, mainly for older people, outlining the levels of automation to be achieved. The final objectives relate to societal needs (road safety, reduction of congestion, service to regions in demographic decline and/or ageing) but also to improving competitiveness.
- The "SIP" (strategic innovation promotion programme) launched in 2014 is an inter-ministerial programme that aims to facilitate the circulation of basic research results to commercial applications. (For information, in addition to autonomous driving systems, there are 11 priority areas: thermal engines, power electronics, energy storage in hydrogen form, underwater mining exploration, infrastructure maintenance, natural disaster detection, cyber security, IT applications in the agricultural and mining sectors, IT applications in manufacturing - we let the reader draw conclusions on these industrial and R&D policy choices and compare them with the French case). The objectives of the ADUS autonomous driving for universal service component are road safety and the reduction of congestion, the deployment of autonomous driving systems, and the creation of public transport systems for the elderly.
In its Phase 1 (2014-18), the SIP-adus programme's industry-academic-research-government R&D themes were: high-precision dynamic mapping, man-machine interface, cyber security, sensors, AI.

Japanese roadmap for the introduction of autonomous driving (2018)

Source: presentation by the National Policy Agency at the UN Economic commission for Europe, march 2019

The organisation chart of the promotion structure of this policy is clearly dominated by executives from manufacturers. Unlike Korea, where program directors are mainly from universities and public institutions, two Toyota executives have successively held the position of SIP-adus program director.

The NEDO (new energy and industrial technology development organization), Japan's largest public research project management organization, coordinates the various ministries and public agencies involved, industry stakeholders, and public research institutions and universities.

SIP-adus Program Promotion Committee
Source: http://en.sip-adus.go.jp/sip/file/sip_en_2018_achievement_s.pdf, 2019

Finally, it is important to stress the "societal" goals of the Japanese programme, which aims to respond to the challenge of the ageing of Japanese society and the announced decline in the population.
- Japan is already the country with the highest share of people over 60 in the world (33%, compared to 29% in Italy or 26% in France). By 2050, this proportion is expected to rise to 42%, slightly above other developed Asian societies and Mediterranean European countries (all above 40%. France would be at 32%)
- In addition, the population peak in Japan was reached in 2010 (128M inhabitants). Since then the population has been declining.
- In addition, as in other developed countries, the very dense Japanese rail network is in the process of shrinking.

It is in this context that various experiments with autonomous shuttles in rural areas are being carried out.

3. The French case

The French case obviously seems more familiar to us; its particular characteristics become more obvious when compared to other countries.

Jeehoon Ki notes a number of plans aimed at various objectives ( revitalising the French industry, "freeing" the economy, accelerating the energy transition, redirecting mobility), of which the development or deployment of the AV is an objective or component, and a national strategy explicitly dedicated to it.

- The AV is one of the 14 items of the NFI (Nouvelle France Industrielle) plan, aimed at reviving French industry around specific ideas and products, and launched in 2013 under the leadership of Arnaud Montebourg, Minister of Productive Recovery. It includes a roadmap for AV development. The plan is " trimmed " in 2015, this second step specifying the R&D objectives. The roadmap provided for a "package" of action plans: coordination of initiatives by DGCIS (Directorate General for Competitiveness, Industry and Services, now DGE), including international cooperation with China and Korea; acceptability studies; launch of R&D projects coordinated by ITE Vedecom and the Ministry of Economy on man-machine interfaces, AI, connectivity; experiments on test roads; adaptation of the regulatory framework (road authorisations, labelling, etc.) by a group comprising the ministries concerned and adaptation of the infrastructure by the DGITM.

- On the side of the Ministry in charge of infrastructure, transport, environment, and energy issues (now MTES - Ministry of Energy and Solidarity Transition), various legislative provisions (Energy Transition Act of 2015), Ordinance of 2016 concerning road tests,

- The AV is also mentioned in the 2019 "PACTE" law, which includes provisions among a range of free-market measures to extend "the scope of AV experimentation in order to supervise more advanced experiments on the national territory, whose priority objective is safety validation on real use cases" ("The possibility of conducting experiments with drivers inattentive or operating the vehicle at a distance (from a remote control centre for example) will be open. The measure will specify the criminal liability regime applicable to these experiments". p.53 www.economie.gouv.fr/files/files/2019/PACTE_Juin2019/bro-a4-pacte.pdf). The explicit objective is to "enable our manufacturers to be leaders in the global race for the number of kilometres covered in autonomous mode".

- A national strategy for the development of the VA is presented in 2018, which coordinates the various administrations and ministries concerned. Jeehoon KI summarizes its objectives as follows: to authorize autonomous public transport services under human supervision from 2020, and level 3 private vehicles; to aim in 2024 to authorize level 4 vehicles on the roads and to develop autonomous solutions for logistics

- Finally, the Loi d'Orientation des Mobiilités / LOM (Mobility Orientation Act) discussed in 2019

A particular feature of France is the visibility of autonomous shuttle experiments. Taking up the results of Fabio Antonnialli's work, Jeehoon KI reminds us that 58% of autonomous shuttle experiments take place in European countries (14% in North America, 16% in Asia). 8 of the 53 European projects take place in France - two of the companies developing this type of vehicle, Navya and EasyMile, are French. The majority of these autonomous shuttle tests are about micro-transit.

Source: Antonnialli (2019)

The table below shows the landscape of autonomous shuttle manufacturers today. However, we can guess that in this matter the landscape is very fluid and a presence at best symbolic today does not say much about future positions...

image 18 - France - constructeurs de navettes

Source: Antonnialli (2019)

4. Summary comparison

Jeehoon KI notes several common points in the policies pursued in the three countries studied:

- France, Korea, Japan, have identified and chosen the development of the AA as an essential element of an industrial competitiveness policy;

- In all three cases, it is the ministry in charge of transport issues that plays the greatest role in policy design, particularly in defining the regulatory framework.

Among the differences, he notes:
- The final objectives of the AV development and deployment: to improve mobility in areas with low public transport service in France, to address the problem of mobility for the elderly and/or rural areas in demographic decline in Japan, and to be more openly industrial and commercial in South Korea;
- The innovation strategy sees the emergence of two opposing models, one demand-oriented ("demand-pull") - France, where the objective is to make VA socially acceptable - the other focused on technological development ("tech-push") - Korea, whose objective is to develop marketable technologies
- Focus (or not) on AV experiments: France and Japan have several autonomous shuttle demonstrators in experimentation. Korea is more focused on testing C-ITS Cooperative Intelligent Transport Systems (ongoing in 5 cities)
- The definition of international standards: France seems more committed to this task than others- Technological choices: Jeehoon Ki notes that France is moving towards connected VAs, with an important role for infrastructure, unlike the Korean strategy.

In support of his demonstration, Jeehoon KI finally gave us a lexicographical study conducted on three important framework texts (France: strategic orientations report of May 2018, 96p.., Korea: Action plan for innovative growth engine of May 2018, Japan: Action plan for realizing automated driving of March 2018): he noted that the Japanese document refers proportionally, much more to "technology", components, and telecommunication technologies (especially 5G) than the French document, in which the terms referring to mobility and social acceptability services (also very present in Japanese documents) are more present.