It has been almost 20 years since the DARPA "SmartDrivingCar" Challenges and 15 years since Google jumped in with the objective of providing demand-responsive automated mobility and we're still "envisioning" as if this challenge is like that of achieving nuclear fusion... always "50 years away, but once we get there, then the return to society is absolutely non-trivial". Presented is my perspective on automated mobility gained from a more than 50 year career largely focused on achieving safe, affordable, equitable, sustainable high-quality mobility for our towns and cities. Reviewed will be what was tried and why it failed; where we've been recently and why it has struggled; and where is it that we might be going and what are the challenges and chances of success. Looking forward, the concept of Autonomous Transit Networks is presented and assessed using the Princeton Synthetic Daily PersonTrip (PSDPT) data set of the 1.1 billion individual person trips made by the 320 million persons on a typical day. The mobility opportunity and operational challenges of several conceptual deployments are presented. The challenges associated with the scaled realization of such deployments will be discussed.

The world has changed considerably for transportation planners since April 2020—and it is still changing, in many ways that are fundamental to how we work, shop, play, seek health care, and travel. The COVID-19 pandemic significantly altered the remote work landscape in the U.S. and there is growing evidence that at least some portion of the remote work trends will stick beyond the pandemic. The next normal for work will result from complex interactions amongst multiple factors, shaped by the experience of both employees and employers whose joint decisions will determine the evolutionary pathway of the hybrid workplace. To help understand the forces shaping these experiences, we have conducted separate longitudinal tracking surveys of both employees (7 waves) and employers (5 waves). We aim to understand employee satisfaction with telework, and reveal and characterize clusters of telework trajectories through and beyond the pandemic. We study the effect of telework on the activity participation behavior of teleworkers, especially with regard to the frequency, location and timing of out-of-home non-work activities. We also study the employer-side perspective on telework from a descriptive analysis of responses from top executives at 100 large employers in North America with regard to how remote work evolved through the pandemic and how employer response to remote work varied across departments, sectors of operation and their pre-COVID and April 2020 remote work policies.

Several factors are converging to drive a new industrial revolution in the U.S and beyond. Advances in manufacturing technologies are reducing costs and lead times as well as creating new industrial capabilities. Companies are rethinking supply chain resilience in the face of global supply chain challenges brought on by the pandemic, climate change and geopolitical crises. And the federal government has passed historic legislation in the past year plus that will lead to billions of public and private-sector investment in key industries such as transportation, semiconductors and clean tech in cities and states across the country. Mindell and Reynolds will discuss the confluence of these developments, their implications for manufacturing, scale up, innovation, and the work of the future and the opportunities they present for the MIT Mobility Initiative and MIT more broadly.

Habit and its potential hysteresis effect, have been recognized as difficult problems in travel demand modelling and forecasting for over 45 years. Past efforts to deal with the problem have included the formulation and estimation of panel data models, including the potential help of shock effects in the series. In the last years, and with the incorporation of techniques borrowed from psychological research, we have been increasingly capable of understanding the role of habit (and also of other latent variables, such as being risk prone or averse or being green or uncaring about the environment), by using hybrid choice models. In this talk, a quick review of these issues will be provided, and also some recent examples of the new approach in the case of modelling bicycle use, car and motorcycle ownership, and motorcycle use.

At the COP in Glasgow, in the Inflation Reduction Act (the US’s largest infrastructure commitment since Eisenhower), in wealth country subsidies, and in major philanthropic investments, the electrification of personal vehicles has been touted as the single most important way to address CO2 emissions in the transport sector. It’s not. I’ll do a scan of life cycle analyses, the implications for lithium supply, demographics of the planet, and the opportunity to multipurpose investments and then discuss better and more effective policy choices. Promoting quality car-independent lifestyles is paramount.

Sequential decision problems are an almost universal problem class, spanning dynamic resource allocation problems, control problems, optimal stopping/buy-sell problems, active learning problems, as well as two-agent games and multiagent problems. Application settings span engineering, the sciences, transportation, health services, medical decision making, energy, e-commerce and finance, but in this talk we will emphasize applications in transportation and logistics. These problems have been addressed in the research literature using a variety of modeling and algorithmic frameworks, including (but not limited to) dynamic programming, stochastic programming, stochastic control, simulation optimization, stochastic search, approximate dynamic programming, reinforcement learning, model predictive control, and even multiarmed bandit problems. We will present a universal modeling framework that can be used for any sequential decision problem in the presence of different sources of uncertainty, using a “model first” strategy that optimizes over policies for making decisions. We will present four (meta)classes of policies that are the foundation of any solution approach that has ever been proposed for a sequential problem, either in the research literature or used in practice (including policies that have not been invented yet). We will close by making the case for teaching sequential decision analytics at both the undergraduate and graduate levels, including to students in fields centered on applications as well as methodology.

A pressing question for deploying autonomous vehicles is: will they be safe enough? The usual answer of "at least as safe as a human driver" (e.g., Positive Risk Balance) is likely to be both too simplistic and much more complex than might be apparent. Which human driver, under what conditions? And are fewer total fatalities OK even if it means more pedestrians die? Who gets to decide what safe enough really means when billions of dollars are on the line? And how will anyone really know the outcome will be as safe as it needs to be when the technology initially deploys without a safety driver? This talk covers risk acceptance frameworks, what people mean by "safe," setting an acceptable safety goal, measuring safety, accounting for uncertainty, a framework for deciding when to deploy, and a sampling of ethical challenges that must be addressed. The emphasis is not on how to build machine learning based systems, but rather on how to measure whether vehicles based on this technology will be acceptably safe for real-world deployment. An approach to deployment safety should include at least: a governance model that encompasses stakeholder concerns, a sound safety case, a definition of acceptable risk, an approach to safety engineering, a metrics strategy, an approach to uncertainty management, a defined deployment decision process, a security plan, a safety culture plan, and a plan to address relevant ethical considerations.

This talk focuses on the disconnect between the incentives for academic researchers and the production of collective wisdom to inform transportation planning and policy. While the former emphasizes numbers of publications and citations, the latter is about producing knowledge in a manner that produces societal good. Unsurprisingly, the emphasis on counts has led to tremendous growth in the number of academic papers. However, it is questionable whether this leads to a commensurate amount of wisdom and is arguably harmful. For example, encapsulating the literature is arduous not only due to the volume but also because uniqueness is paramount for publication. To that end, thoughts are presented on how incentives may be realigned to increase the societal impact of academic research in travel demand, and input from the audience is also solicited. Included in the talk will be research related to autonomous vehicles, telecommuting, and public transportation.

After a century of car-oriented urban growth, cities around the world are implementing policies and plans that aim to make their neighborhoods and streets more walk-able and transit oriented. Renewed attention to non-auto mobility is driven simultaneously by the impending climate crisis, public health concerns, and inter-city economic competition. This has given rise to increasing demands to shift the dominant research and policy focus from merely electrifying, sharing, or automating the vehicle fleet, to developing city environments that are more conducive to pedestrians and transit riders to begin with. This session invites three active mobility experts to explore what it takes to achieve more walk-able and transit-oriented urban environments from the perspectives of public policy, current research, and history.

Dr. Matthias Winkenbach, Director of the MIT Megacity Logistics Lab, will discuss how novel technologies and advanced algorithms can enable future last-mile logistics services to create superior value for consumers and businesses. Based on his lab’s research with major players from the supply chain and logistics industry, he will reflect on the disruptive potential of recent advancements in machine learning and artificial intelligence, as well as technology innovations such as drones. He will share his perspective on how one can harness emerging technologies to create the next generation of last-mile logistics services by connecting the power of algorithms with the wealth of human experience.

Advanced Air Mobility (AAM) operations—characterized by electric and hybrid aircraft, and highly-automated or even autonomous operations—are expected to dramatically transform the way in which we transport people and goods. The deployment of these new vehicle types, business models, and aircraft operators will increase competition for already constrained airspace resources. In addition, AAM will also require a shift in how traffic management services are provided. In this talk, I will discuss emerging AAM traffic management challenges, and some of our initial work in overcoming them.

Prof. Zhao and MIT Transit Lab researchers will showcase three sets of A.I. applications in public transportation: prediction, monitoring, and control. 1) Prediction: deep hybrid model with satellite images and graph embedded road networks for demand prediction in Chicago; 2) Monitoring: computer vision for bus travel time estimation in Boston and text mining for sentiment analysis in WMATA; and 3) RL-based bus operation control in CTA. We will cover examples of unsupervised learning, supervised learning, and reinforcement learning and give a flavor of some future research: causal analysis with ML, generative AI, and multi-channel view of cities.

Innovation and Entrepreneurship (I&E) often go together. Following the 19th Century product innovations of Karl Benz, the most important auto industry innovator/entrepreneurs in the 20th Century were probably Henry Ford and Eiji Toyoda/Taichi Ohno, noted primarily for process innovation. Although Ferdinand Porsche built and sold the first-known hybrid-electric vehicles in 1901, it wasn’t until the late 1990's that Toyota reintroduced hybrids to then relatively slow moving auto industry, and triggered a soft re-introduction to the world of electro-mobility. In this century, Tesla has catalyzed dramatically increased clockspeeds across the industry, with (at recent count ) over $850B committed to EV development and production during the remainder of this decade. Electrification has reignited I&E across the automotive and transportation value chains. As companies grow and mature, their I&E muscles can begin to atrophy, as young startups race ahead, typically unencumbered by the organizational complexities inherent in the industrial giants. As such the automotive industry in this decade seems to be populated with numerous large, incumbent auto makers trying valiantly to re-ignite dormant I&E cultures within their walls, racing against the dozens or hundreds of startups, ready to take risk in search of outsized rewards. One domain of some interest is in the development of public EV charging networks. Europe seems well underway in building out a viable EV charging network. In the U.S., Tesla’s still-mostly-closed network works well for its customers, but many non-Tesla EV owners find that state of public charging to be quite poor, limiting the market to those who have single family homes or other situations where they can get assured charger access. In this talk, we will share some thoughts on the challenges facing the auto industry’s transition to EV’s, with particular emphasis on the dynamics of I&E in large and small players and — possible paths to scaling public charging in the U.S.