Cleaner cities: What technology is needed for smart waste bins

We all know this problem: The local glass or paper container down the road is full again, but according to its schedule, the waste management company will only be back next week thursday. The result: People are stacking their trash next to the bins. Unprotected by weather, the wind blows the stuff over the street, glass is being scattered all around the containers. This trash is most likely lost for recycling. Not only it makes our cities look dirty, the municipality is losing money. Intelligent waste management with smart containers could avoid that. Harald Naumann, engineer in information electronics, explains how.

While it is possible for waste management companies to calculate their routes and cleaning intervals according to experience and seasonal factors, i.e. festivities like christmas or local events, modern IoT technology could help optimizing the cleanup process significantly. This would not only save the company a lot of money, it would also help in avoiding overfilled bins. Finally, each kilogram of glass or paper that is not lost for recycling means additional money for the municipality as well.

In this article, I’ll focus on the city of Hanover, in the state of lower saxony north western germany, and its urban hinterland. In 2014, 24,283 tons (24,283,000 kg) of waste glass was collected for recycling. This corresponds to approximately 22kg per resident. (Source: Aha, the local recycling company for region Hanover.) The mean annual price for recycled glass increased from €36.2 per 1000 kg in 2000 to €48.3 in 2008 and subsequently decreased to €45.3 per 1000 kg in 2011. (Source: Eurostat. http://ec.europa.eu/eurostat/home.)

If Hanover was able to persuade its residents to increase the amount of glass that they recycled by just 20% then the city could raise an estimated €220,000 per year. Within five years, it could raise an additional million Euro more.

Some facts on region Hanover:

• Hanover region residents = 1,100,000
• Hanover city residents = 523,000
• Hanover region area = 2,291 km2
• Hanover city area = 204 km2
• Waste collection points (waste glass, paper, old clothes) = 660

A waste collection point is a combination of at least three containers for waste glass (white, green, brown colour) and two for waste paper and sometimes further containers for old clothes.

• One collection point will support 1667 residents
• 660 collection points means an estimated 660 x 5 containers on average = 3300 containers
• Hannover city has 313 waste collection points = 1900 containers
• Approximately half of the inhabitants of the region live in the city of Hannover

We can assume that trips to containers that are not yet full can be saved, but to do that the waste management will need to know in advance, which these are. And that means sensors. The technology for detecting waste level in a container is well established and practical but to add value to the use case information about each container needs to be captured and transmitted on a regular periodic basis to a central location where the data can be analysed to provide the basis for a collection truck route plan. This means wireless connectivity – a good, reliable, cost-effective, low power, wide area (LPWA) network technology.

Technical needs

Full level sensors at waste glass containers will send one status update message of 10 bytes once per day; some might send once a week depending on use with the location generated by GPS.
In Hannover Suedstadt we can already see 15 waste bin collection points per square kilometre. A dynamic route calculation for the trucks could help to save truck-roll, diesel, wear and tear and depreciation on city assets, as well as free up local authority employees for more productive work. In addition, an efficient collection service leads to a more pleasant user experience motivating a greater proportion of residents to fall into line with the social benefits of recycling. Nobody likes to visit an overfilled waste glass container.

Two years ago, the volume of the standard domestic waste bins in Germany was reduced. In return, they provided homeowners with free of charge bins for the waste paper. In parallel, a private recycling company offered homeowners another free of charge waste bin for the paper. Waste paper, it seems, is big business – they are competing to get their hands on it! Moreover, each of them run their own trucks to collect it; every second week two trucks arrive in my small dead-end street to collect the waste paper. If two companies are hot to get our waste paper, then it must make sense to collect more paper and glass in the inner city of Hanover.

So let’s consider how each of the primary LPWAN technologies might address this real-world requirement.

NB-IoT

For NB-IoT the one message per day limitation is no problem because it has no duty cycle restriction and is easily able to transmit more data per day. However, the waste bins in Hanover run on a battery. Intelligent design would eliminate unnecessary full level updates or GPS positions. S/W updates over the air are also possible with NB-IoT. We can expect a very high downlink communication speeds at with NB-IoT using a mode called multi-tone to use 12 NB-IoT carriers in parallel.

SIGFOX

SIGFOX can do the job. SIGFOX technology offers a maximum of just 6 messages per hour only. In this case, we need one message per day only. However, SIGFOX cannot offer downlink for firmware updates because they are limited to 4 messages in download by 8 bytes only.

LoRaWAN

LoRaWAN can do the job as well. Full level transmissions are uplink only. LoRaWAN does not offer true bi-directional capability and upload and download are on simply old Aloha protocol without Listen before Talk. A high package loss can be expected. LoRaWAN offers no broadcast capability. All nodes need to be connected on a case by case basis. The LoRaWAN network will be busy with downloads during an end device downlink transmission and not able to receive any uplink message from any node during an SW update. LoRaWAN cannot receive a message during transmit; it offers half duplex only. Firmware updates in the field are not possible.

Weightless

Weightless technology uses TDMA with listen-before-talk (LBT) and Frequency Hopping. LBT and Frequency Hopping mean that Weightless has no limitation on duty cycle. It is like LoRaWAN on the license free band. You will save the subscription fee for the SIM card of euro 2 for 3300 waste containers per year.

How to deploy the technology

Let us now turn our attention to an example deployment in a typical small to medium sized German city like Hanover. Weightless will avoid the subscription fee required for a connection to a SIGFOX or NB-IoT network. Moreover, Weightless avoids the subscription fee for other devices in a smart city like smart parking or street light. Hanover has an area of only 204 square kilometres. The TV tower called Telemax is 282 metre high. The old TV tower called Telemortix is 141 meters high. One Weightless gateway could cover all waste bins in the inner city.

However, the wider urban hinterland around Hanover has a size of 2,291 square kilometres. For this reason, a hybrid public and private network service provision make sense. In small cities one Weightless gateway will be enough to cover the whole city whist waste bins in small villages around the larger cities, a smaller number of households that might not be able to justify their own network could be covered by NB-IoT.

Weightless technology also offers true bi-directional communications capability-enabling FOTA because, in addition, Weightless uses time division duplex. TDD enables both uplink and downlink transmissions using the same channels to avoid package loss.

Conclusion

The winner for this waste container example is Weightless technology, followed by NB-IoT or a combination of Weightless and NB-IoT. SIGFOX technology and LoRaWAN are okay if the operator is happy to accept the risk of a deployment that is not future-proof with a firmware upgrade capability. Maybe recycling collection points do not constitute critical infrastructure, at least from a security point of view so firmware upgrades are not as important in this use case as they would be in, for example, a traffic management application. And 3000 end devices located in positions that are routinely visited on a weekly basis are not impossible to upgrade manually. Whereas end devices in less convenient locations, or in cases where there are considerably more end devices in a network could certainly not be economically serviced by regular physical visits to upgrade software.

In a similar version this article had been originally posted in the IoT M2M Blog.