Cohort 3 AgriFoRwArdS student, Paul-David Zuercher shares his experience on travelling to Singapore to present as collaborating author the paper “Optimising virtual reality training in industry using crowdsourcing” at the 12th Conference of Learning Factories, held at the prestigious Singapore Institute of Manufacturing Technology.

I’m Paul, a computer scientist studying robotics and autonomous systems and researching with the University of Cambridge. Due to my research training grant, I had the opportunity to attend the Conference on Learning Factories in Singapore, and now have the privilege to report about it.

As a researcher, I feel the need to contextualise my narrative with some Background information. This report will not only share insights about the lessons learned in my field of expertise – human technology convergence – rather I hope to share some interesting insights and thoughts about visiting a foreign country as a research ambassador, the role of autonomy in agriculture, the impact of technical autonomy on humans, and the fauna, flora, and culture in Singapore. I share many experiences I will value for as long as I can remember them. For our pleasure – my dear reader – I’ll add some interpretations and nuances to prevent us from drying out as I was thirsty for both of us when walking through the hot streets of Singapore.

I love learning about cultures! Living and interacting with institutions and societies around the world. While these interactions are joyful on their own, I also find pleasure in reflecting on cultural norms and conventions to break free from our cradled cultural perspective to increase our ability to appreciate humanity’s diverse body of knowledge, arts, structures and strivings humans develop under different circumstances. My journey to appreciate cultures in a systematic manner started with my two years as the head of PR at Engineers Without Borders (Germany, Darmstadt). At Engineers Without Borders, we had many intercultural collaborations with for example Kenya and Tanzania.

I soon realised that intercultural communication is a key aspect for making a difference in a globalised world, and first conceived how intercultural communication can be academically analysed in Engineers without boarders’ intercultural communication workshop. Since then, I read many non-fiction books about this subject that proved incredibly valuable in intercultural communication. I count myself privileged to have received that education which proved critical when I started studying in a foreign country (UK). Today, my thoughts are deeply grounded in cultural theories and models, and I hope sharing this perspective with you – my dear reader – proves valuable to you too.

However, I’m not only observing human-made culture; with my professional work, and my research at the University of Cambridge I also contribute to human creations. My expertise is centred around computing, autonomous systems, human technology convergence, and machine learning. Such a contribution is a recent paper I co-authored with Thomas Bohné, Vera Maria Eger, and Felix Mueller: “Optimising virtual reality training in industry using crowdsourcing” and published with the Conference on Learning Factories was is the reason for my travel to Singapore. Okay, but enough background information; let’s start with our Journey!

Singapore – a journey worth the costs?

The natural habitat of 4.54 million Singaporeans is GMT+8, more specifically a small (about 700 km^2) island below Malaysia and above Thailand called Singapore or “lion city” that lies one degree north of the equator. That means two things: Firstly, the plan of my sleeping pattern according to the latest research insights to minimise jet lack (for details see “Why we sleep”) was a hopeless endeavour since I only stayed in Singapore’s time-distant position for four days. Secondly, the 4.5 hours of delay in my connection after Istanbul lead to an overall transit time of 18h.

Luckily, I met a young student from Switzerland during the waiting time who was waiting for the same connecting flight. But the time spent at the airport had no chance of being noticed since we lost track of time in engaging discussions, background stories, and ambitions. She is currently studying sustainable management practices with a focus on east Asia shipping. She told me that she would learn a lot about designing sustainable logistics – more specifically, aiming to minimise the impact of shipping on the environment. We were tired but I didn’t omit the elephant in the room: “Will you ask your program directors how sustainable it is for all members to fly to Singapore?”. “No, but it is an invaluable experience to visit the logistics industry in person to understand the gap between theory and practice”. People are not free of bias. And I often ask myself how differentiated we have to think to get rid of bias. Is it even possible in such invaluable situations? Is there anything we can get away from this situation?

Judging it as irresponsible to fly to Singapore seems to play into the human tendency to quickly judge and overestimate the quality of own opinions over – maybe – a carefully curated underlying strategy of sustainability. Maybe when considering the new contacts she is creating in Singapore, the industry visits she gets inspired by and the drive she is getting through the experience leading to a greater impact on her career; I didn’t create the logistics sustainability syllabus – maybe if I did, I  would come to the conclusion that it would be irresponsible to miss the learning opportunities in Singapore; missed lessons that will amplify to massive opportunity costs by changing her future working to a worse performance due to lacking insights or even to a career change.

Just ignoring the extensive amount of CO2 (about 3723.88 kg per person) required to travel to Singapore, seems to be irresponsible too. Unfortunately, I don’t have an answer to this question. Maybe awareness of the trade-offs made, combined with trust in the individual’s expertise and critical thinking mediates the potential bias of judging the situation. I’m not an expert on climate change but give my best to be aware and raise awareness of the impacts of my own actions. I gave my best, telling many people I met that it was very affordable (only approximately 6.8% of the overall cost) to offset emissions. How effective are these solutions in counteracting losses? – another very hard-to-judge question^[1]. On my trip, I learned a lot and hope sharing these insights with you will further increase the trip’s positive impact on our global society.

Arriving in Singapore

The four days I spent in Singapore certainly only equipped me with a naïve understanding of the island. The touristic exploration of the adjacent and popular attractions constrains the horizon of my per- and conceptions. Most of my time with researchers from the Singapore Institute of Manufacturing Technology (SIMTech), is under the jurisdiction of Singapore’s government Agency for Science, Technology and Research (A-STAR). Thus, my experience cannot represent such a diverse city-state as Singapore – how could I, if even critical thinking researchers report about their challenges to portray the complex historical evidence about the young (founded in 1965) in history books? However, I was impressed by the people I had the privilege to talk with! I experienced excellent technologic transparency regarding employed systems and obtained results and witnessedtheir industry partner’s readiness to provide access to their internal information about challenging digital transformations. A supporter of the conference – who shared the most fantastic restaurant with me; thank you if you read this! – contextualised this open knowledge policy as Singapore’s economic strategy. She said, Singapore is too small to position itself as a leader in the producing industry, so it strives to be a knowledge hub and interface between systems and cultures.

Besides the exchange with academics, I also enjoyed the diversity of Singapore’s vibrant flora, vivid fauna and culture everywhere. The vegetation is of such radiating colour, I had to touch a few plants to be sure they are not artificial. No plastic indeed, just a carefully curated beautiful diversity of flora and fauna. Due to the intense urbanisation of the small island, Singapore lost 95% of their natural forests over time. While in the 1960s Singapore’s lands only added up to 550 km², astonishing human-environment engineering projects reclaimed an additional 130 km² by 2015 and are expected to claim additional land. A minimum of 10% of land overall is dedicated to parks and nature reserves, with increasing efforts to preserve the remaining wildlife. Singapore’s landscape is marbled with spots of natural refuge in form of parks and reserves. With the ambitious goal to position a park within 10 minutes of walking distance wherever one strolls around.

UNESCO proclaimed World Heritage status to Singapore’s Botanic Gardens in 2015 and remains the only Botanic Garden that is World Heritage to date. The gardens earned this status due to their “Exhibit an important interchange of human values, over a span of time or within a cultural area of the world, on developments in architecture or technology, monumental arts, town planning or landscape design” (criterion ii) and “Be an outstanding example of a type of building, architectural or technological ensemble or landscape which illustrates (a) significant stage(s) in human history”. How architecture, technology, and nature is best exemplified by their recent addition of 10 super trees in a new botanical garden. These “supertrees” (see image below) are vertical gardens that produce solar electricity, serve as air vents for surrounding conservatories, and catch rainfall. The high-tech nature-technology-hybrids generate electricity through integrated photovoltaic systems to deliver energy to power the park and cool the park down through convection.

Communication cultures

At the latest, since the first day I arrived in the UK – looking right, left, right and nearly getting hit by a car – (I was raised in Germany, where cars come from the other side) I’m very cautious how I behave in new cultures. I first spend a few days analysing, synthesising and adapting to the layers of the metaphorical onion of social conventions, professional conduct and personal values. As a guest in this beautiful country and generous conference, I wanted to appreciate their efforts and express my gratitude on all cultural levels. To my astonishment, I noticed no systematic discrepancies between the cultural values of the SIMTech researchers and the cultural heritage I appreciate in the unique working environment at Cambridge. People that are aware of the deep entanglement between British colonialism and Singapore might have a starting point for understanding the relationship between British and Singapore cultural values. I enjoyed many intercultural knowledge exchanges. They reminded me of the constructive process of knowledge creation I often enjoy with my Cambridge supervisor Thomas Bohné. I think both of us are trying to minimize our “opinions” in order to maximize our collaborative knowledge creation through differentiation and critical thinking. In my fruitful discussion with “George” Zhao – the highly talented and brilliant principal researcher of SIMTech’s impressive model factory – our discussions often converged to exceptional and carefully curated constructive insights.

For me, an intrinsically valuable argument is valid (relies on true assumptions) and sound (combines the basic assumptions logically). People that can think logically can still argue about the truth of statements if their assumptions are not coherent (e.g.,they don’t share the definition of a used word). In social argumentation even things that are not logically sound can also be regarded as sound arguments if they convince people – nevertheless, in scientific discussions, that follow the classic analysis of knowledge, establishing shared assumptions is crucial for constructive discussions. However, for scientists, it is often challenging to dig down to the essence of what makes them disagree. Often, it’s less mentally demanding to just say things such as “No.”, “Yes, but…”, “You should”, “We must”, “This is not correct.”. If I hear those words – either because I say them or because I sit in the same room that frequently uses constructs like these, I often become worried about the opportunity costs of all knowledge getting lost when discourse fails to integrate the unique knowledge and experiences of panellists. Maybe the style of discussion reveals your perspective on what is valuable – With George, I never worried about losing insights. Even though George is much more experienced than I am, we gave each other space to present our insights and carefully curated the differentiated position of all arguments and supported each other in a joyful synthesis of new insights by integrating our collective knowledge. But instead of listing our dialogue, I’ll integrate the insights from the many interesting discussions I had the privilege to be part of.

Guidance from leading academics

Prof Hummel – a coryphaei (one could say mother) of learning factories and in her long career it could be reasonable to assume that she had all thoughts I had and much more. I asked her about her perspective on the position of humans in the industry. She showed me her perspective and raised lots of interesting questions and research challenges. She guided the discussion, constraining arguments based on the immense expertise and knowledge she possesses to lead me to a point where we think – I say we, because she was kind enough to hold my hands the whole time until we arrived there – the position humans should have in highly intelligent systems. She is an important and (thus?) busy woman and I highly value the time she took to share her wisdom with me. This exchange will certainly help shape and motivate my future research.

First, challenges are centred around dynamic workflows. Especially for small lot sizes, it’s often not beneficial to automate the processes of producing and assembling them. Analogously, for research in the agricultural sector, the process of grabbing and handling different fruits and performing tasks like pruning effectively is complex to achieve with robots but simple to achieve for humans. Combining the strength of both and seeing how they are. A major shared challenge is thus to create robots and autonomous systems that are highly adaptable. Second, we must rethink our approach to the optimisation of manufacturing processes.

Her inspiring determination to align any further progress towards shared collaboration – in which humans and robots decide together what should be done – raises the hope that we are shaping the future of robotics and autonomous systems towards truly human-centric industries. Third, Prof Hummel emphasises the necessity for humans to be at the centre of future factory designs. I still have questions but certainly appreciate every second with her and am thrilled to see that she – the now vice-president of the international association of learning factories – envisions the future of the manufacturing industry in enabling equal collaboration between human and autonomous systems (e.g. robots).

Prof Ramsauer’s brief outline about sustainability in manufacturing factories and he told me about his most recent plan to help policymakers reduce industrial CO2 emissions by on average 15 per cent in an economical way (just by implementing automatic switch of mechanisms when machines are not in use (e.g. using light barriers).

Prof Hammer – who I asked about how we can best overcome the challenges of integrating economic and ecologic considerations in practical research – emphasized that the major difference between ecology and economy is that ecology has constraints. Since economic considerations only make sense constrained by the ecologic capacities of our ecosystem, thus rendering ecologic balance sheets as the starting point for any economically sustainable research project.

Dr Zhao Yizhi (who is so nice to also hear by the name “George” – a close phenonym that is easier pronounceable for many people) and his team, presented us with their brilliant dashboard, that enables fast investigation of key performance in real-time. The KPIs are separated for investigations on different layers. He proudly reports the “wow” factor he often achieves with this multilayered system. The first layer is the supply chain – which presents the most important KPIs for multi-site logistics -, the second level is the enterprise’s site – which shows the aggregated KPIs of all sites shop floors – and the third level is the shopfloor – which presents collated information about the floor’s individual machine KPIs. This interface allows top managers to “drill down” suboptimal performant components of sites, floors and machines to enable making adequate interventions in the system in order to overcome inefficiencies.

George proudly emphasizes that the in-house build tool is interactive, allowing one to identify the exact problem while only requiring a few inputs from a manager. Maybe by now, the question “how can we identify the KPIs?” crossed your mind too. George and his team talked to many managers and collected much data from Singaporean companies to identify the critical information they need to know.  With the tremendous resources of SIMTech and the marvellous data access they have through government programs; George’s combination of observational studies and interviews with numerous managers that are complemented by investigating relevant data is evidently an excellent method to identify the most crucial KPIs.

What is the relationship to your research, Paul?

On the technology side, my research covers the spectrum of solutions that enable interaction between humans and robots. To support the vision of effective Human and Robot teamwork, on the human side, we need to display more relevant information to the human (e.g. simple concepts such as guiding attention and complex concepts such as personalised systems that adapt to the user’s knowledge) and build interfaces that supply higher quality input about the persons state better (e.g. simple metrics such as stress level and more complex such as expertise and satisfaction). On the robot side, robots require simpler, human-centric inputs to allow managing the robots towards valuable objectives, while humans still need to be able to take over controls in dynamic environments. Regarding the outputs of robots, we require more transparent and human-centric decision-making procedures and means of communication.

Let us quickly talk about humans and technology. Are you familiar with technological determinism? Technological determinism accredits inherent social development trajectories to technologies. From this perspective, the invention of smartphones that constantly accompany and inform us naturally leads to people looking down on their phones more. Ultimately, the work of us researchers who develop Human-Robot interaction systems will determine whether Human-Robot interaction is technology-focused where the robot imperatives humans or whether humans keep their autonomy and collaborate with autonomous systems. In which future do you want to live? ☺

People who are familiar with agricultural practices probably also identified parallels between the levels of analysis in Industry 4.0 and the levels important to the agricultural sector. Many agricultural companies have responsibilities in the supply chain to communicate how much agricultural product they can supply to their business partners; they have multiple sites which they need to manage with collective resources, and they need to know the state of each field to make sure their crop survives. Better performance would certainly be possible and quantifiable when providing key performance indicators and “drill down” methods are available to agronomical decision-makers.

So What?

Okay, so what? Let’s briefly contextualise and integrate these insights. The concepts of Industry 4.0 and beyond are about making better decisions by collecting data and optimising workflows using smart systems. Determining which components of the current processes limit the overall performance is reflected is indicated by the key performance metrics. They can help to reveal what needs to be tweaked to improve the overall performance. In manufacturing systems, the most important key performance indicator for an enterprise with costly machines is overall resource efficiency. In some systems improvements of one per cent can make a difference of millions of dollars. However, the key performance metric is not relevant on every level.

Dedicated time reflecting with George about solutions for the agriculture sector

The current lack of data in the Agronomic sector, renders many industries 4.0 solutions designed for non-volatile factories impractical for the agricultural sector. Later, in a more personal discussion, George and I exchanged some alternative ideas and opportunities of how the KPIs could be identified – especially when resources are low, and data is sparse. We discussed an idea I had during his showcase, to conduct a flow analysis of all components necessary to deliver the agricultural product with the goal of analysing and optimizing the entire process. While this is quite a mouth full, it is based on a common method to optimise processes: value stream mapping. In brief, value stream mapping consists of (1) analysing and linking all processes with their in – and outputs; (2) identifying the bottleneck that prevents a higher performance. Subsequently, it’s only necessary to synthesise the most critical components (i.e. bottlenecks) of the linked processes (imaginable as a “web” or graph). Some of these will not be changeable (e.g. setting the growth time of strawberries to zero is pretty much impossible). However, other process components can be influenced to increase performance (such as optimising soil components or using robots for specific process components). Of course just because it would theoretically make sense to improve something does not mean it is economically and ecologically reasonable.

Therefore, these opportunities need to be analysed for economic viability (and – latest since our experience with Prof. Hammer we know this requires – including an ecologic balance analysis). Inter alia, allowing to identify opportunities to increase the performance of agricultural practices, by drilling down from high-level KPIs to the exact process components that limit the agricultural performance. Furthermore, this perspective allows retrofitting which (combination of) robotic systems would be economically and sustainable and how they could be used to optimise the agricultural system. These metrics could even be quantified and the time necessary until you would have a positive return on investment (amortisation) could be determined. George’s face, initially sceptical, got more and more excited and he shared much more valuable information with me that will guide my quest to create better human-centric tools towards a resilient global and national supply chain of agriculture by quantifying opportunities for optimising agricultural processes.

George and I talked a little bit more about the challenges and opportunities of AI and simulation to design more effective industry 4.0 solutions. And we dreamed about a time when there is no need to find the relevant information to optimise your system, but it is a time when the optimal solution finds you. A world in which every worker is immersed with an interface displaying the most relevant information personalised for each worker (based on e.g. expertise, position, and the state of the system) – ( just as I dreamed with Prof. Hummel about a future) with workers being autonomous in choosing what they want to do based on their individual motivation and autonomous systems taking care of tasks which are not satisfying or too monotonous for them.

I left the conference with many new ideas on how we can create a sustainable future. Hopefully, sharing my experiences and insights sparked a similar synaptic explosion in you as the conference visit did for me💡

If you have any questions or if you want to discuss anything further, I’m looking forward to hearing from you: pdz20@cam.ac.uk. And if you want to keep updated on my research with articles like this and my papers, you can subscribe to my publications over on my website: pauldavidzuercher.com.

^[1] https://www.ecowatch.com/best-carbon-offset-programs-2650926204.html

A huge thank you to Paul for allowing us to share his thoughts and experiences travelling to and attending the CLM2022 conference in Singapore. Having the chance to learn from others in this way and to expand our knowledge helps us become more thoughtful, deeper and more collaborative thinkers and researchers.

Do you want to get involved in the CDT?

If you are interested in learning more about the breadth of what we do and if you share a passion for agriculture and technology then go to our AgriFoRwArdS CDT website to see more about our research, how you can be involved and how to apply to be a student in the program.