The Charioteers

To people familiar with Indian philosophy, the title may bring forth images of a divine chariot described in the greatest poem on war. This story though is less about that One Charioteer, and more about the charioteers we meet every day, and those we will need tomorrow. This is a study on the evolution of technology and business models in mobility; and on what it is all about — the people involved. But now that the image is formed, let’s linger in the metaphors a little longer …

We are all travellers, never intending to stay put for too long. The journey through time and interstellar space is involuntary and inexorable. Whereas, the travels in our own spatial frame are usually by choice, and timebound. There is often a purpose to the journeys we undertake, but there are some where there is no other purpose than the journey itself. The destination is pre-determined for some travels, and there are those where we are willing to meander and let destiny decide. In all these journeys, it is the charioteer who guides us safely to our destination, ready for the next one.

In the journey of life, our body forms the chariot, and our intellect the charioteer. For the more mundane travels, we have had animal powered and wheeled transport from the early days of history. As these evolved into chariots and carriages, we have each learnt how to navigate with them through our journeys. And yet there has always been a need for trained, professional charioteers — either dedicated, or for-hire. Charioteers who can guide us through the journey, affording us the safety and the reliability of their experience.

Other modes of transport have also evolved — finned boats and ships; winged aircrafts; and land-based mass transit systems, from camel caravans to steam locomotives and further. All these modes have an even greater reliance on experienced guides to steer them safely and efficiently to their destination.

Our modern-day chariot — the automobile — is a marvel of 20th century engineering that continues to evolve. Yet the role of the charioteer, aka driver or chauffeur, remains largely the same. They are our guides in the journey, navigating us through the maze of city roads, the pulsating traffic, and the ever changing weather.

In the remainder of this article, we will evaluate trends in mobility i.e. how people move from point to point; with particular focus on how shared mobility will evolve, and how it may affect the role of the charioteer.

Shared Mobility

Over the last century, shared mobility was mostly in the form of taxicabs. Taxicab, shortened as taxi or cab, is a portmanteau of taximeter — the device used to measure the time and distance that the vehicle has moved; and cabriolet — a light horse-drawn vehicle. The service associated with hailing cabs on-demand is termed ‘ride-hailing’. The three primary service models were street-hailing (pickup set at rider location), taxi-stand hailing (pickup set at driver location) and pre-booking (pickup set at booked location). Besides ride-hailing, the other on-demand services are long-distance booking, and fixed-duration booking.

Pre-booked rides were serviced by dispatching a vehicle from the nearest taxi station at the desired time. Radio-taxis enabled a more efficient dispatch by using radio communication with mobile drivers. With the epitome of technological innovation and integration — the smartphone — pre-booking is now available as an app-based ‘e-hailing’ service. Connected phones have made it possible to hail a ride with the push of a button, and to substitute for the taximeter using their inbuilt GPS. A plethora of service models have been enabled with this technology.

Businesses offering on-demand ‘ride-hailing’ service are referred to as ‘transportation network companies’ — these ‘aggregators’ pool the services of fleets and individual drivers onto a common platform. This includes services like Uber, Lyft and Ola. This service is also referred to as ‘ride-sourcing’ in the sense that it outsources ride requests to the drivers on the platform. These companies may also offer ‘ride-splitting’ services wherein multiple riders can split a ride — this includes services like UberPool and Ola Share.

The service model for pre-arranged pooling of rides is referred to as ‘ride-sharing’. This includes one-sided matching where service providers try to pool together a number of riders e.g. airport shuttle; and two-sided matching where a matching platform enables peer-to-peer pooling e.g. BlaBlaCar for intercity travel.

The emerging service model associated with renting a car or bike for short periods of time, referred to as ‘car-sharing’ or ‘bike-sharing’, is set to revolutionize private ownership of vehicles. Businesses that are seeding this model includes Zipcar, Savaari, and Zoomcar for cars and ofo for bikes. If a private vehicle can be picked-up and dropped-off on-demand, the motivations for private ownership disappear. The added benefit of being able to choose the type, make and model of the vehicle makes the service model even more attractive.

Along with enabling the various service models, smartphones also serve as an assistant to the charioteers, providing them information on navigation routes, and traffic and weather conditions. In all these models, the core responsibility of ferrying people safely and comfortably to their destination is still squarely on the charioteer.

Challenges in Mobility

The chief problem with urban mobility today is inefficiency. Across the world, private vehicles occupy most of the available street space and parking space, and yet serve only a third of all trips. These vehicles lie idle for at least 90% of the time, and the parking space required to accommodate them takes up to 20% of urban real estate. Despite this, people spend a large amount of their driving time stuck in congestion — up to 26% in Moscow, and 6% in Dallas. To add to this, on average, 30% of the cars in congested downtown traffic are simply cruising for parking.

Perhaps counter-intuitively, the answer to these challenges of urban mobility does not lie in wider roads or more parking space. After all, the goal of mobility is to move people and goods — not the vehicles themselves. Wider roads and convenient parking will only serve to incentivize more private ownership. The solution should aim to reduce private vehicles, and improve the utilization of vehicles, roads and parking spaces. This the approach adopted by the various Shared Mobility options. These options harness technology to bring their utility value to be at par with private vehicles.

The key problem that Shared Mobility attempts to solve today is the problem of operational efficiency. Below is a list of some of the key processes required to achieve this.

1. Operational — these are the activities required before rides can be available

• Demand Prediction — in order to determine the required supply at a given time and place, there needs to be an accurate prediction of the demand. These may be based on real-life factors such as weather changes, sports events, concerts, holidays, elections, school-breaks etc.
• Capacity Management — in order to service the demand, the appropriate capacity has to be enrolled, and deployed at the right time and place. Although demand can be influenced, there are more tools available to control the supply.
• Pricing Strategy — this a key tool that can be used to both influence the demand, and to control the supply. The pricing strategy aims to maximize earnings for the ride-provider without exploiting the rider, while complying with regulatory caps.

2. Pre-ride — the goal here is to optimally match a rider to a ride-provider.

• Positioning — accurately determine the on-road locations of the rider and the available rides
• Routing — determine the best path between the rider and the available rides, taking road restrictions into account
• Distance and Time estimation — calculate the distance and travel time for various routing alternatives
• Matching — match a ride-provider to a rider based on various requirements and preferences. The match may be in real-time, or it may be scheduled for a later time. Matching is significantly more complex in case of pooled rides, where riders are matched to already running trips.

3. On-ride — the goal here is to ensure that the ride is safe and efficient

• Tracking — accurately track the location of the rider and ride-provider. The tracking needs to be resilient to disruptions in network and location availability. Monitor these and other factors for any anomalies, and prepare to respond if required.
• Adapting — adapt to changing traffic and weather conditions to suggest the best route for the trip
• Engaging — provide appropriate information updates to the rider and the ride-provider for the duration of the trip.

4. Post-ride — the focus here is to build a lasting relationship

• Fare Calculation — although the fare estimation is often done upfront, the actual fare needs to be determined after the trip, as the route or destination may have been changed.
• Rewards and Referrals — customized rewards, and referral options are critical to building a lasting relationship and to take advantage of network effects
• Continuous Improvement — use rider and ride-provider feedback to provide actionable insights for continuous improvement

Each of these phases and processes are opportunities for mobility service providers to differentiate themselves, and offer a better service to riders and ride-providers.

Some of the key technologies that have evolved to enable these solutions are:

• mobile data services — the ability of a device to connect to the internet while it is mobile. A breadth of such services are offered by numerous telcos around the world.
• geolocation services — the ability to determine the precise geographic location of a mobile device. Such services are offered by a number of government owned satellite systems. The most readily accessible of these is the United States’ Global Positioning System.
• geographic information systems (GIS) — the ability to create, manage, analyze and view geographical data and drive decision making using the data. Google Maps is the most readily accessible GIS platform.
• cloud IT services — the ability to setup a reliable, secure and scalable IT infrastructure accessible over the internet. Such services are offered by a host of companies, the largest being Amazon (AWS), Microsoft (Azure) and Google (GCP).

Each of these services have sufficiently matured to the point where they can be used to create viable solutions to the above problems. Reliable on-demand shared mobility requires all these functions work consistently for most of the time. However, the solutions need to be designed to handle fluctuations as well as temporary outages in these services. While these services themselves evolve to offer greater functionality with higher reliability, mobility services providers also continue to innovate for a better quality of service.

Increasing Mobility

The development of vast road networks since the end of the second world war has greatly reduced the frictions in human mobility. This reduced friction has led to greater ease in the exchange of ideas, goods and services. In turn this has led to overall improvement in global productivity and prosperity. The increased prosperity further drives an increased demand for mobility, forming a virtuous cycle.

Urbanization and globalization may have their supporters, and their critics, but the endeavor for prosperity is everlasting. Over the next decades as productivity and prosperity continue to improve, our desire and need to travel will only increase. To address this increasing demand, transportation will continue to evolve to be safer, and greener.

Safer Mobility

The largest cause of accidents involving vehicles on the road is human error. In US, around 35,000 fatalities occur due to road accidents per year, and in India, there are around 150,000 fatalities. Over 90% of these fatalities are caused by human errors such as speeding, or miscalculating other drivers’ behaviors. The primary purpose of Autonomous Vehicles (AV) is to reduce fatalities by replacing human drivers with intelligent machines.

Let’s review some of the key characteristics of human drivers. Drivers need to be equipped with the knowledge of how vehicles work, what the traffic rules are, and have some basic knowledge of the lay of the land. Then, they need to be skilled to simultaneously watch and listen, and perform different controls with each of their limbs, while constantly analyzing and anticipating their surroundings. Broken down into the individual stages, we realize the complexity involved in the simple act of driving a vehicle.

Given the complexity, human drivers are surprisingly safe. But there are accidents, and these are largely due to human errors. One of the largest sources of such errors is distractions, such as the use of mobile phones, leading to a lack of attention to the surroundings. Then, there are times when the driver may be incapable of paying attention due drowsiness, or due to the influence of alcohol or drugs. There are those that voluntarily drive recklessly in violation of traffic rules. And then there are times when drivers lack the skill to navigate adverse weather or road conditions, or vehicle malfunction.

AVs need to firstly be able to drive like a human, perhaps in limited situations, and then be able to drive more safely than humans. The SAE has defined a number of levels of autonomy, depending the extent to which AVs address the challenges described above:

Levels of Autonomy

As autonomy proceeds from Level 1 to Level 5, the level of Artificial Intelligence (AI) required also increases. Various Machine Learning (ML) algorithms are used to implement these different levels of autonomy, covering the three robotic primitives:

1. Sensing — this is essentially a classification problem in machine learning.
2. Planning — this typically involves reinforcement learning — one of the approaches to ML.
3. Acting — this is typically broken down into further sense/plan/act stages hierarchically.

A lot of the progress in driving autonomy are due to the advancements in ML algorithms in solving the so-called ‘Weak AI’ problems. These advancements derive from the improved capability to capture and store much larger amounts of data, and the improved capability to process these large amounts of data. However, even with all the advancements, current technology does not perform accurately 100% of the time. Occasional errors in voice recognition by the latest sequence-to-sequence approach will be acceptable to most users. Such errors though, will not be acceptable when it comes to self-driving vehicles. Recent fatal accidents are testament to the open technological challenges. It is also clear though, that within the constrained environments, autonomous driving can be safer than humans.

Over the next few decades, we will see a higher adoption of autonomous modes in specific driving scenarios e.g. highways, parking lots, traffic jams. Given the rapid pace of evolution in technology, and increasing urban congestion, this is both desirable and inevitable. However, there will continue to be a need for trained drivers for all the remaining scenarios that will be encountered in any transit. This is particularly required in urban environments, more so for the chaotic and disorganized traffic encountered in developing countries.

Looking further out, perhaps multiple decades, technology will have to find solutions for the much harder ‘Strong AI’ problems before we can approach Level 5 autonomy. Once this level of autonomy is possible, transportation will be significantly differently from what what it is today. A number of companies are investing in creating fleets of ‘Shared Autonomous Vehicles’ (SAV).

This will be different mode of transportation that can combine the benefits of private vehicles (private space) with public transit (managed vehicles on public guideways). This evolution of this mode of transportation will also require a broad acceptance of moving away from private ownership. Given the higher costs of building AVs, this will likely be the preferred mode of deploying them once Level 5 autonomy is truly achieved.

Greener Mobility

Greener mobility refers to the constant improvement in travel efficiency in terms of reducing travel time for lower monetary and environmental costs.

About 20% of all CO2 emissions are from road transport, and about 15% is contributed by passenger vehicles. As emission standards around the world continue to clamp down on the permissible limits, Internal Combustion Vehicles (ICV) will no longer be viable and the only option will be Plug-in Electric Vehicles (PEV). PEVs use electricity from the grid as their source of energy, and are able to store the energy in rechargeable batteries. PEVs include battery electric vehicles (BEV) and plug-in hybrid electric vehicles (PHEV). For BEVs, the grid is the sole source of energy, whereas PHEVs also use an on-board consumable fuel as an additional source of energy.

In order to evaluate the overall benefit of PEVs to the environment, we need to consider the well-to-wheels (WTW) environmental cost. This includes greenhouse gas (GHG) emissions from the manufacturing, and the operation of the vehicle, as well as the GHG emissions from the process of providing energy to the grid. This analysis shows that in most developed countries, PEVs have a significantly lower WTW GHG emission compared to all other forms of transport. However, in developing countries, the usage of fossil fuels implies that there is no reduction in overall emission due to PEVs. In order to realize the benefits of the shift to PEVs, there has to be a shift to renewable sources of energy such as wind and solar. The shift to PEVs is however a precondition to reduce the dependency on fossil fuels. PEVs can also become an integral part of the renewable energy infrastructure with the deployment of Vehicle-to-Grid (V2G) charging points.

The evolution of battery technology is central to the deployment of PEVs. The battery is the key component of PEVs in terms of cost, weight and life, and its replenishment infrastructure is the largest factor in their operational efficiency.

For BEVs, the battery capacity ranges from around 15 kWh for the Mitsubishi i-MIEV, to 100 kWh for the high end Tesla Model S. EVs require between 150 Wh and 250 Wh per kilometer depending on vehicle weight, speed and terrain. This can be seen in the driving range for the i-MIEV of around 100 km, and for the Model S, around 500 km, as per the US-EPA test procedures.

There are multiple options available for the battery replenishment infrastructure i.e. domesting charging, public charging stations and exchange stations. The charging times for the batteries depends on the power supply used.

• A single phase domestic chargepoint can supply up to 3.3 kW at 16A and 7.4 kW at 32A — leading to a 6 hr or 3 hr charge time for 20 kWh (~100 km).
• Three phase domestic or public charge-points can supply between 11 kW and 43 kW at 400V for charge times between 2 hrs and 30 mins.
• External fast DC chargers at charging stations can supply up to 120 kW, reducing the charge time to 10 mins.

For personal vehicles that can be kept off-service for periods of time, with small batteries, charging stations can be more cost effective. Whereas public vehicles that need to be constantly in service with larger batteries, exchange stations will likely be more time effective. Exchange stations also allows business models where the battery is leased separately instead of bundling it with the vehicle, reducing the initial cost for buyers.

The price of batteries has been falling from $1000 per kWh in 2010, to around $200/kWh today, and is expected to fall to $100/kWh by 2025. Today, the battery contributes around half the cost of an EV, pushing the overall price of EVs to 150% of a similar ICV. By 2025, it is expected that the battery will contribute around 25% of the EV cost, bringing the total vehicle cost at par with ICVs. However, in terms of total cost of ownership (TCO), EVs are already at par in some regions due to the lower operational cost, and more significantly, due to government subsidies and incentives.

Looking to the future, the drive by governments to reduce dependency on imported fossil fuels, and their impact to the environment, we are likely to see a significant shift to EVs over the next decade. Given that TCO is already at par for EVs, fleet owners and ride-sourcing companies will find it viable to operate EV based intra-city services in the near future. As EVs become prevalent, drivers will shift from being one of the main agents of pollution, to becoming the champions of environmental protection.

Infrastructure for Mobility

The need for human interaction and mobility lies at the heart of urbanization. As mobility evolves, we will also see an evolution of the urban landscape. Historically, the layout of a metropolis has been determined by the means of transport available for people to commute. European cities are more dense as they have been built based around a walking commute. On the other hand, newer American cities became more decentralized because of the availability of cars and road networks.

The design and layout of public transit systems has been central to urban planning, as people tend to cluster around their stoppage points. Urban planning includes the layout of roads, and the layout of residential and business areas. Land use zoning has been used to delineate types of activity, and enable efficient public and private transit between the zones. Various alternate approaches to urban planning have emerged, balancing the energy efficiency gained from a compact urban form with the broader quality-of-life aspects gained from the dispersed city. Form-based code uses the physical form of structures as the organizing principle instead of the separation of use — this gives regulators the means to achieve development objectives with greater certainty. With a transit oriented development (TOD) approach, urban infrastructure is built to be within walking distance of public transport. A mobility oriented development (MOD) approach offers a more flexible alternative, allowing multiple modes of transport with a wider reach for public transport.

Along with urban transit, as a result of increasing domestic and international economic integration, the demand for long distance travel is also increasing. Long distance transit may be by land, sea or air; can be served by public or private transport systems; and can be used to transport people or goods. A broad network of national and state highways have been developed across the world since the end of the second world war. These roads have been a crucial component of economic growth and are used by public bus services, as well as numerous private alternatives. The terminal points of long distance travel, such as airports, railway stations, and bus terminals, also become the hubs of a city’s urban transit system. The transportation system of the future will allow seamless transition across various modes of transport to allow point-to-point mobility across the globe.

As mobility become safer and greener, roads will transform from physical spaces to intelligent systems. The infrastructure for mobility will need to evolve like other intelligent networks such as communication networks and energy networks. This Intelligent Transportation System (ITS) has its own ability to sense, plan and act. ITS systems use cameras and other sensors to determine the state of vehicles and roads. They then use this information to plan various courses of action based on objectives such as law enforcement, prioritizing emergency vehicles, and ensuring the smooth flow of traffic. This system relies on communications networks for sending and receiving information. The ITS capabilities will continue to improve with improving sensing, processing and communicating capabilities. 5G communication is expected to make a significant impact with the possibility of vehicle-to-vehicle (V2V) communication enabling controls like signal-free intersections, and vehicle platoons on highways.

The broad deployment of shared autonomous vehicles will be accompanied with a complete shift in how cities are designed. These vehicles will need lesser road space as they will be able to drive closer together. Traffic rules that are designed for humans can be simplified as the vehicles will be able to communicate among themselves. It may also be possible to eliminate private ownership and parking spaces altogether.

Regulations and Mobility

The intrinsic nature of the need to travel effectively makes mobility a public utility. Any glitch or disruption in mobility services can have an adverse effect on large segments of populations, particularly in urban areas. Transportation laws have been legislated across the world to set the rules for different modes of transport. These laws vary across countries, sometimes drastically so, with variations across states and even smaller jurisdictions. The laws have also evolved over time alongside other changes in transportation.

Regulations for on-demand vehicles are intended to regulate the market so that it is fair for all — owners, drivers and riders. Various historical issues that regulations have tried to address include:

• Undersupply of cabs leading to increased wait-times, and overcharging
• Oversupply of cabs leading to reduced driver earnings, congestion, and fall in quality
• Exploitation of riders e.g. uncertain fares, price fixing, circuitous routing, ride refusals
• Exploitation of drivers e.g. underpaying by owners, violent competition among operators
• Unintended consequences of regulations such as corruption in giving licenses, and lack of innovation in operational efficiency

Largely, rider demand is a function of waiting time and fare, as they look to travel from point to point. Similarly, the driver supply is also a function of dead mileage and fare, as they look to maximize their earnings. So the goal of any regulation is to minimize waiting time for rider and drivers, and to allow the fare to find the right equilibrium. The key variable that can be regulated are the fare rules, and the supply of cabs — with variations depending on time or zone.

As a public utility, there need to be guidelines around safety and quality of service for various modes of shared mobility. Service providers also need to ensure that due taxes are collected and paid, and that appropriate insurance is taken for drivers and riders. E-hailing applications are able to use technology to monitor and control quality, taxes and insurance. However, safety assurance has been found to be lacking at times. Regulations need to ensure that drivers and riders are appropriately held accountable for safety. And as always, the charioteer lies at the core of ensuring that the transportation system is safe.

With the growing importance of personal data in a breadth of digital services we use everyday, there is an increasing realization of the need to ensure the data is not mishandled. In Aug 2017, a supreme court bench in India ruled that the Right to Privacy is a fundamental right. The European Union made the General Data Protection Regulation applicable to all member states from May 2018 — the key objective being to give users control of their personal data. Mobility service providers need to ensure that data security is built into their solution, and that personally identifiable information is kept secure at all times.

As autonomous vehicles evolve, mobility regulations will also need to adapt. Some cities around the world have allowed testing of driverless vehicles. There is typically a condition that there will either be a safety driver, or the self-driving car company must monitor and be able to take over driving remotely. There are also places that have disallowed driverless vehicles for concern that it will eliminate jobs.

There is a genuine concern that the jobs of drivers may be eliminated once autonomous vehicles are broadly deployed. This is part of a broader concern about the impact of automation on human employment. While computerisation has been historically confined to routine tasks involving explicit rule based activities, artificial-intelligence allows it to enter the domain of pattern recognition based non-routine cognitive tasks. In order to prepare for this shift in human jobs, public policy needs to invest in developing uniquely human skills such as creative and social skills. Policy at a local or national level can be used to slow down the pace of technological advance. However, technology advances at a global level, and it will continue to evolve to bring greater efficiency to all economic activities. So it is important for public policy to also invest in developing technologies such as autonomous vehicles with the goal of ensuring an equitable distribution of its benefits.

Future of Mobility

As seen in other industries, the highest value in a market accrues to platforms that bring together the providers and consumers of a service. In mobility, this is apparent with the immense value generated by technology enabled ride-hailing services. The goal of these platforms will be to provide end-to-end transportation by linking together different modes to enable Mobily-as-a-Service (MaaS). If mobiles phones created the ‘sharing economy’, MaaS will create the next economic wave that some call the ‘passenger economy’ worth $7 trillion.

This shift will be characterized by freed up time and improved productivity. MaaS will displace vehicle ownership, and create a new landscape of services covering human mobility, logistics, and long-haul transportation. MaaS will be delivered by businesses with fleets of autonomous electric vehicles and bring together multiples modes of transport. Consumers will have their means of connecting with these businesses and will see mobility as a service no different from the network service provider they use for connectivity. And we will perhaps go back to being our own charioteers.

This article also appears on LinkedIn.