[Read more session reports and live updates from the UNCTAD E-commerce Week]
The event, co-organised by UNECE and CEFACT, discussed the latest technological innovations in the field of artificial intelligence (AI), autonomous things, quantum computing, and edge devices with three main focuses on data source, data communication, and data computation.
The event was introduced by Ms Maria Rosaria Ceccarelli (Director-in-Charge, UNECE Trade Division), who underlined how the Internet in just one generation - no older than thirty years - has changed our private and professional lives. Technology is changing the world and it is necessary to understand the challenges that new technologies create. Additional introductory remarks were sent by Mr Vint Cerf (Chief Internet Evangelist, Google), who argued that the topics taken up in the session will feature the next ten years of telecom and computing evolution, requiring new forms of leadership to engage and set vigorous and ethical developments.
The first part of the event was moderated by Mr Todd Frazier (FedEx) and focused on the latest technology trends in data sources, trying to answer the question of whether we will have completely autonomous vehicles.
Mr Davide Scaramuzza (Professor of Robotics and Perception, University of Zurich) talked about autonomous flying drones. In order to understand the technology, it is useful to conceptualise how current drones work. A first type of drones can be characterised by remote control that requires a line of sight or communication link and a skilled pilot; a second type is features a GPS navigator on the drone to provide position measurement for the autonomous drones. However, sometimes the signal at high altitudes can be inaccurate. The question is then how to localise drones without GPS. The closest technique to GPS is to replicate GPS using Ultra-Wide-Band Localisation (UWB). An alternative technique is based on Lidar-based Localisation, already developed in self-driving cars. The shortcomings include high costs and high weight. As a result, it is impossible to install it in a small drone. Another technology that can be utilised is Vision-Based Localisation, which with a single camera allows the drone Visual Simultaneous Localisation and Mapping (VSLAM). Nonetheless, as the panellist argued, autonomous navigation is not solved: the use of drones without GPS are still limited to contexts with a lot of light; in addition, the open challenges for Vision-based drones are related to perception: as a result, current studies are trying to use machine learning to complement the models. Finally, looking at the future, the panellist explained that a likely application of drones will see reconfigurable drones, able to change their shape, used in rescue contexts after disasters.
Mr Jörg Jermann (Senior Consultant, Rapp AG) gave a presentation on Autonomous Driving. He explained automation as a five stage development of autonomy, described as follows: no assistance, assisted, partially automated, highly automated, fully automated, and autonomous. The stages from partially automated to fully automated represent a shift of responsibility from the driver to the car. Nonetheless, a distinction should be highlighted between the notion of autonomous driving, meant as a form of absence of interaction with the environment (the technology is able to read the environment but not interact with it), and automated driving, in which the car behaves in the optimal way considering the information received in a ‘send and receive’ exchange of data. Going into the specificities of the technology, the panellist highlighted two elements to take into consideration: sensors and data. Sensors represent a one-way form of communication, whereas data represent a two-way form of communication that requires digital safety and security. This results in different levels of connection:
Finally, with regard to possible strategies, the panellist highlighted two possible approaches: the evolutionary approach, meant to scale stage by stage; and the revolutionary approach, starting with fully automated vehicles in limited contexts. In both approaches, the challenges are connected to security and control, but most importantly to the type of approval procedure by authorities, complicated in the case of a car being only software. As a result, the commissioning of each automation level will be in the responsibility of the car manufacturers.
Mr Francois Guichard (Mechanical Engineer, UNECE Transport) talked about the regulatory aspects of Intelligent Transport Systems and Automated and Connected Vehicles. He recalled the work of the World Forum for Harmonization of Vehicle Regulation (WP.29) in the establishment of detailed regulatory provisions. WIth regard to the regulatory challenges that the Forum is facing, the panellist highlighted the goal of achieving a balanced moderate proactive approach in the promotion of new technology. The aim is to integrate technologies into existing transport systems and to ensure that the benefits of such technologies can be captured. Current achievements have been reached in the area of safety of automation derivatives, in the assessment of environmental performance of vehicles, and in the resolution of issues related to cybersecurity and data protection. On the last point, the draft regulation on ‘Cybersecurity and Software Update’ (OTA) should be recalled.
Mr David Woudenberg (Data Scientist and Lead Autonomous Technician, Xomnia B.V.) presented Xomnia’s pathway from a self-driving boat to autonomous shipping. The company, based on big data projects, developed autonomous shipping boats starting from current studies on self-driving cars and translating them into driving boats, which then became shipping technology.
The second part of the event was moderated by Mr Jan Hoffmann (Chief, Trade Logistics Branch, DTL, UNCTAD) and focused on the latest technology trends in data communication.
Mr Laurent Vieira de Mello (Head of Strategy & Planning, Astrocast) talked about the communication of the global Internet of Things (IoT) with nanosatellites. The panellist started his presentation explaining the meaning of ‘new space’ as a new phenomenon characterised by high innovation and the emergence of private spacecraft research and innovation. It represents a smallsat revolution featured by the creation of small size and lightweight satellites, often used in constellation. The work of Astrocast is based on nanosatellites, ground models, and data plans. The use cases that he presented, related to international trade, are applicable in agriculture, fishing, and mining industries. Indeed, global benefits of such technologies are represented by more efficient and cost-effective operations, the reduction of a carbon footprint, accurate tracking of assets, and detailed visualisation of consumption. Nonetheless, he presented challenges with regard to access to best communication frequencies, data security, and specificities of solutions required by different customers.
Mr Nikolai Vassiliev (Chief Terrestrial Services Department in the International Telecommunication Union (ITU)) talked about 5G Networks as the future of the digital communication infrastructure, as shown by its potential in enhancing mobile broadband, massive machine types of communication, and in providing ultra-reliable and low-latency communications. The use of 5G is extremely expensive, yet useful applications of 5G can be applied to IoT such as smart packaging, energy savings, and water control; as well as in ultra-reliable and low-latency communications.
Mr Steven Capell (Australian Government Department of Home Affairs) addressed the topic of API and programmable web. He explained that the last decade has been characterised by the emergence of web platforms such as Amazon and Alibaba, which have challenged the traditional way of understanding the exchange of data. Indeed, the way the web works today and data exchange takes place is increasingly about a web of linked data. This represents a new way of discovering and aggregating data: the conceptual difference relies on platforms in the web talking and discovering each other. It changes the model of data exchange. However, standards are missing with regard to many emerging data pipelines and cloud-based ethernet ring protection switching, which are based on different application programming interfaces.
The third part of the event was moderated by Mr Kaushik Srinivasan (eMudhra) and focused on the latest technology trends in data computation.
Mr Federico Carminati (Chief Innovation Officer, CERN Openlab) talked about quantum computing as a technology that allows the conduction of extremely large amounts of computations. Nonetheless, the use of qubits faces difficulties related to the fact that it is possible to retrieve only one quantum state at a time and it is impossible exactly to copy the states; as well. There is a constant presence of quantum decoherence. Finally, he underlined the role of CERN Openlab at the forefront of research in quantum computing, meant to represent the next revolution in ICTs.
Mr Marc Stampfli (Country Sales Manager, NVIDIA Switzerland) addressed the rise of modern AI as a result of three main driving factors: the immense amount of data available, the rediscovery of artificial neural networks as parallel computing algorithms for deep learning, and the availability of GPU computers capable of enabling artificial intelligence.
Mr Masamichi Tanaka (Chief Strategy Officer, Uhuru Corp) talked about edge computing and data exchange, stressing that the question is not what, but where, edge computing is taking place. Stressing the limitation of cloud computing in managing the increasing amount of data, edge computing represents a mechanism of data elaboration and analysis at the source of the data. Current challenges have been raised with regard to security, data protection, costs, computer power, and latency, so that in the future, issues related to reliability and identification of the elements in the data exchange model need to be addressed. The UN can have an influential role in this.
By Stefania Grottola