Fast, reliable, cheap, and integrated transportation system works as a backbone for economic activities in any city. Due to paucity of good public transportation system, the vicious cycle of owning more cars/personal vehicles and thus congesting cities even more has a terrible effect on the long term competitiveness. Cities like Bengaluru are notoriously known for terrible transportation infrastructure with average peak traffic speed being 4km/hr in some areas like Outer Ring Road (ORR) or Silk Board. Moreover, the situation is not even better for other times of the day - owing to my own personal experience of travelling from Bellandur to Ulsoor (18km) on a Sunday night in 50 minutes. As a result, in cities like Bengaluru, anecdotes of families/couples living apart in different parts of the city to escape the daily office-commuting hassle is not unheard of! 
Both India and China are touted to be a major engine of growth for Asian century. However, compared to Indian growth story, Chinese development has a stark difference of impeccable infrastructure development. Despite the collective failure of engineering, planning, policy-making, design, science and management, there are indeed some good news. Indian government is undertaking massive metro rail development in many Indian cities. Moreover, some common strategies to tackle traffic congestion like more buying buses, adding extra lanes, underpass/overbridge are always welcomed in Indian context. Some of the Indian cities have also encouraged strategies like work from home, tolling, odd-even rule but to little effect. These capacity addition or demand management strategies are very important and should be carried out.
Yet there are more things a city can do without much change in operational budget. Now-a-days, it is not difficult to make data-driven approach to the city’s transportation system. In my view, using modern day computation powered by data rich machine learning based solutions ~ loosely called AI will be instrumental in smoothing of the traffic. I outlay a few cost effective strategies below:
-
Variable Working Hours - According to traffic flow theory, even a 10% reduction in traffic volume can reclaim substantial normalcy in the traffic speed.
For example, in the above picture, at 8000 vehicle/hr, there is a complete breakdown of speed thus forming a bottleneck. However when the volume decreases by ~12% (7000 vehicle/hr), the speed goes back to 90 km/hr. So, taking a cue from this fact, Indian cities should look into lowering traffic flow in the peak hours. The first strategy is to nudge/mandate companies to adopt flexible work hours. Furthermore, a coordinated effort on city scale to collect anonymized origin/destination (for ex: business park or residential society information to maintain privacy), and their flex work hours needs to be carried out. After taking account of road & transportation options, pairwise travel time, it is possible to optimize the schedule of each person and/or companies such that there are minimal overlap of traffic. Such efforts have been carried out successfully as Spitsmijden in The Netherlands. To encourage more and more people joining this initiative, there needs to be individual incentives like daily app-based notifications, expected travel time in odd hours information, vouchers, lotteries etc. -
Bus network redesign - BMTC runs a staggering 2000 different routes with large headway (time difference between two consecutive buses of same route), often intending to reach expensive to serve places. This results in low frequency of important and major routes. Some of the routes like 600 is 117 km long and takes a whopping 5 hours to complete.
Bengaluru Bus Network Longest Route Globally, to increase ridership and towards having a viable bus networks, few of the cities are redesigning their bus routes. The inherent philosophy is to have more transfers (between buses), increase last mile mobility (to serve far areas), higher frequency and optimize for the ridership.
Houston, Auckland, Dublin are few cities which already implemented this overhaul in their system and saw more people using the same without much operational change.
Before Network Redesign After Network Redesign So, what really Bengaluru (or any other Indian City) can do to make bus network more viable? Simplify the bus routes into smaller, more frequent routes that encourages on transfers for longer commute. Anonymized origin-destinations commute time/volume data through ridership,surveys,mobile network,employer records should be helpful in determining routes of importance, demand forecast and therefore ~ adjusting real time frequency.
-
Double down on First/Last Mile options - The average commuting distance of 69% Indians lies in the range of 0-5 km. In the light of high commuting time by car/buses, micro-mobility options like Bounce, Rapido, Yulus of the world have sprung up and expanding fast. These options are great alternatives or compliment to the available transportation options. The thesis that more mobility options ease commuting woes may not be true due to non-sharing of data on common platform as users are locked in. This is a real danger of closely gatekeeping the data through monopoly seeking business behavior understandably due to their business interest. Therefore, like some of the Global cities, it should be mandatory to open the API as a part of permitting/licensing process. Like GTFS data sharing (which enables real time bus/metro scheduling information), it should be mandatory to publish General Bikeshare Feed Information (GBFS). This will make these services much predictable, integrated and possibly enable frictionless payment via One Nation One Card system. The dashboard could look like below:
Open APIs help cities improve their mobility infrastructure, define transit accessibility, help combine trips across different modes or even be helpful for vendors to understand true demands. To kickstart this for Indian cities, I have decided to start a platform for aggregating data from Indian service providers that will fuse data from other sources as well, eventually taking a form of Urban OS. -
EV infra setup & rolling out the same- NITI Aayog has been a big proponent of shared-electric mobility options as a way forward. It is imperative that policies play a very important role in faster adoption of electric mobility - which is a stated goal of Government of India by 2030. Yet, technology - AI in particular - can play a great role in accelerating the country towards EV adoption. India is a very cost sensitive market and it appears that buses and 2 wheelers will be the first to be electrified. Therefore, by detection of unique EV load signature, it is possible to offer targeted incentive for time of use rates or EV specific tariff. Through spatial clustering and demand forecast, optimal charging locations to satiate charging anxiety and revenue maximization can be found out. Neural network can be trained using vehicle information, weather, driving style and road condition dataset for accurate battery management system. Lastly, fleet systems can use city brain (mentioned below) based on battery charge levels for their dispatch strategies. (Ex - To charge in the afternoon at lower rates, ridematching etc.)
-
A coordinated comprehensive city brain - It is a pity that despite higher internet penetration, siloed dataset combined with lack of coordination among service providers (both public and private) results in underperforming transportation resources. Cities like Jinan, Hangzhou have collaborated with ridesharing, IT, transportation agencies to come up with a big data dashboard that gives a city level real time view.
Above is the Alibaba’s version of city brain which promises to provide a suit of transportation product to concern stakeholders. For example - Using dataset of road network, ride order, GPS trajectory, traffic record, weather and User profile, it is possible to train deep neural network to facilitate route planning, estimated time of arrival, optimal pickup-drop location, supply-demand forecast, capacity management, carpooling, emergency dispatch among many other services. Moreover, AI can also help in interacting with infrastructure such as smart traffic light, automated parking-toll payment, dynamic rate determination etc. So, I see no reason why Indian cities will not be able to replicate same level of operational efficiency with some expenditure.
Overall, these strategies needs to work in tandem with massive investment in infrastructure to keep the competitiveness of the Indian urban centers high. It is not that the present problems are unsolvable but the political willingness and urgency needs to be really at war-footing to tackle this issue. Always remember - We have a billion potential problem solvers for our billion problems!