The company uses AI and Machine Learning to automate the production and maintenance of indoor maps and to ensure that these maps match with the actual terrain.
MazeMap, an indoor positioning and wayfinding company, provides indoor maps for universities, hospitals and hotels with large campuses. Founded in 2013 in Trondheim, Norway, the company has offices in Palo Alto, California, London and Melbourne. MazeMap uses AI and Machine Learning to automate the production and maintenance of indoor maps and to ensure that the map always matches with the terrain. The company wants to help people find what they are looking for indoors, such as a room, a moving asset or available resources.
Before the advent of GPS, navigation certainly wasn’t easy. Now with GPS making a foray into indoors in the form of wayfinding, indoor maps are increasingly being used for visualization of information that interoperates different data. The way a building or campus is used, its maintenance can be carried out based on actual usage.
Visualization of information and creating services for end-users is essential when it comes to indoor maps. Be it a smart building or a smart campus, indoor mapping is indispensable.
Because of so many use cases, sometimes multiple maps have to be maintained. MazeMap tries to be accommodative towards its customers and partners so that they can leverage the platform for all use cases, thus minimizing or eliminating any problem that occurs.
Mazemap has a large network of clients and partners, with Cisco being the largest partner providing indoor positioning services and resellers like Dimension Data/NTT and Optus. The indoor mapping company has also joined hands with market vertical-specific solutions like Scientia for universities and other facility management companies.
The company supports the following market verticals: universities, hospitals, corporate offices, hotels, and conference venues. Based on use cases, integrations and update frequencies of maps, there is a difference between all the verticals. For instance, a good user experience for hospitals requires an efficient kiosk interface and integration with the patient notification system, while for corporate offices integration with the meeting booking system and existing employee apps are crucial. Similarly, for conference venues, the layout can change several times a week. This means that even for the same map, usage and integration usually varies.
Asset tracking – seeing where high-value assets are – or solutions like ‘Find-My-Friends’ showing in real-time where your friends or colleagues are, or showing analysis of movements and specialized maps for evacuation etc, are some of the emerging trends in indoor positioning, along with navigation for the blind and replacing physical signage, says Thomas Jelle, CEO, MazeMap, in an exclusive interview. Excerpts.
Indoor mapping relies on GIS and highly precise location, so would developments like real-time location rendering herald new innovations in this domain as well?
Today, there exist solutions for doing indoor positioning down to 5-10 cm. In some use cases, this is important. For example, in determining whether the equipment in a hospital has been through a dirty zone or not (and thereby needing disinfection). The challenge is that many of the solutions are quite costly to deploy – with more scalable indoor positioning technologies I expect new services to arise and a wider deployment of existing use cases.
Tell us about some of the use cases of your indoor mapping platform?
One of the hospitals we work with is St. Olavs University Hospital which previously had five services where different indoor maps were needed. The sad thing was that none of the maps were maintained so the map and terrain wouldn’t match, causing frustration among the users. Today, St. Olavs uses MazeMap as the map foundation for all its services. To be able to serve the different use cases we need to have flexible and powerful APIs. As such, APIs are at the core of our strategy.
With hundreds of use cases, we would have no chance of solving all of them even with a thousand developers, so we pick some use cases that we want to solve using our resources. We build APIs so that customers and partners can build on top of our platform to solve their particular use cases. Some customers only use MazeMap for wayfinding, but more and more are starting to use our APIs. Let’s take the case of the University of Twente, where they have used our platform to visualize heaps of stuff – they have one service showing available parking lots in real-time, another service showing the temperature in all rooms, and a third one showing congestion of people in real-time.
Also Read: 5 key Location Intelligence trends for 2019
Enhanced data analytics capabilities and expanding Cloud capacities have opened new avenues in location intelligence. How do you think the indoor mapping segment can leverage it?
We see that analytics are already seen as important by our customers. Both real estate owners and facility managers deploy it to monitor their usage of buildings and also for presenting the data to the end-users. This enables them to easily see the availability of rooms, desks, find quite zones, or know about the queue in cafeterias etc.
Cloud capabilities mean faster roll-out and update frequencies on new features and improvements. The main challenge with cloud solutions is privacy and GDPR, as the building owner might have less control over the data in the cloud or so it might appear, even if this is not the case in reality.
By 2022 the global indoor location market is expected to touch $40 billion. Amidst this manifold growth, what are some of the major challenges that the industry faces?
There are a few obstacles in the path of growth in my opinion. One of them is security. Some organizations are reluctant about presenting indoor maps digitally to users because of a perceived security risk – what if the maps get into the hands of a person with harmful intentions? In most cases, these concerns are not legitimate because all buildings have fire and escape maps that are public in the building. Also, when you create indoor maps for end-user usage you can exclude areas that are sensitive. There is also an option to protect digital maps with usernames and passwords, making them much harder to access then the printed maps in a building.
Another challenge I want to highlight is that the industry is much segmented. We count about a hundred players that have entered the market within the last 3 years, including new start-ups. Some of these players offer really good and sustainable solutions, but others are still throwing spaghetti at the wall to see what sticks. The latter are creating a lot of extra work for customers, generating skepticism towards the industry and making them hesitant to try other solutions.
Another challenge is to solve a lot of use case integrations that are needed – many of the systems that hold data you want to visualize on a map doesn’t have APIs or have very complex APIs, creating costly and time-consuming integrations.
The last challenge I want to mention is a basic challenge that few of the existing vendors seem to have good solutions for but is crucial for the long-term success in the market. Every year between 5-10 % of the buildings changes totally. For the terrain and the map to align, the maps need to be updated when the changes occur. Without automation, it requires a lot of manual efforts to keep the maps up to date in a university or a hospital with a large campus. At MazeMap, we have automated this approach by integrating with vendors of facility management solutions, so when a floorplan is changed it is reflected in the end-user maps as well.
Artificial Intelligence, IoT and technological convergence like Geo.AI will reorient location technology paradigm. What do you think would be the impact of these emerging technologies on indoor mapping?
AI can help capture and keep the maps up to date as buildings change. We use AI to build and maintain the maps. But AI can also be used for predictive usage of buildings and understanding how buildings are being used. In combination with sensors, this can make the building smarter in terms of being more efficient and more adapted to the user needs. Examples are lower energy usage with heating and cooling, and a comfier indoor environment with different temperature zones in the building. In all, these technologies are drivers for the need for indoor maps as many of these smart building features require some form of visualization.
What is Mazemap doing for more sophisticated predictive models and visualization tools?
From the visualization side, we do quite a lot. For example, showing available parking lots and meeting rooms. We visualize if meeting rooms are available and integrate with sensors to see if the room is actually in use or if it is just booked and not utilized. If it is booked and vacant, we wait for some time to check if someone occupies the room, but after a certain time limit has passed we cancel the booking and make the room available for others. Other visualizations we do are things like heatmaps, available desks, temperature etc. to let users make decisions regarding where to work etc.
How do you ensure that Mazemap supports multiple information sources and is often easy to integrate with other APIs and systems?
Indoor maps are a basic infrastructure with hundreds of use cases. We are not able to develop solutions to all these use cases, but we still want our customers and partners to be able to use our maps. In order to support this, we spend a lot of resources developing flexible and powerful APIs. As I said before, APIs are at the core of our strategy. We build APIs so that customers and partners can build things on top of our platform to suit their use cases.