More droid trouble

Above: 77 new realtime radiation monitors made by JB Japan Brand were installed in Fukushima in March 2015 and taken offline weeks later. This unit, in front of Odaka Station in Minamisoma, has an “under maintenance ” notice posted on it.

Back in December of 2012 we posted a two-part article which took a close look at problems concerning the government’s realtime radiation monitoring posts which dot Fukushima and nearby prefectures, which we dubbed “droids.” At the time we pointed out that about 3000 of these devices, built by several manufacturers, had been installed, and we highlighted problems with the hardware and its placement, the online maps the government provides to inform citizens, and the procurement process. In particular, the system was enmeshed in a scandal then involving a smaller manufacturer, Alpha Tsushin, whose units were inadequately tested before being installed and were subsequently taken offline, and the company’s contract rescinded. We’re amazed to see a nearly identical scandal involving new droids from a different manufacturer unfold again.


Above: Exterior and interior views of four superficially similar monitoring units from different manufacturers. Compared to the more reliable units by NEC and Fuji Electric, the JB Japan Brand droids look poorly designed and built, even possibly incomplete. Units built by Alpha Tsushin were also taken offline soon after being installed in 2012 due to reliability problems.

Several weeks ago, Safecaster Joe Moross noticed some new realtime monitoring posts in Fukushima, similar in external appearance to those made by Fuji Electric, but made by a relatively obscure company, JB Japan Brand. Joe took a “close” look, and what he saw did not inspire confidence. Shortly afterward, we began to hear rumors that these units were having big problems, and then that they had suddenly been taken offline. Since then there have been several articles about the problems in local and national newspapers and some informed commentary on blogs and on Twitter.

To summarize the situation, JB Japan Brand , a small Fukushima-based company which has sold other radiation-detection equipment under its brand, won a 52 million yen bid from Fukushima Prefecture in late 2014 to provide 77 new monitoring posts to be installed in sites in seven municipalities in Fukushima. They were installed by the end of March 2015, and put online on a test basis. On March 30, the Japan NRC, which is responsible for hosting the data on its online maps, determined that 13 of the 77 units were not transmitting data. They informed the responsible person at Fukushima Pref. by email, but this person apparently did not inform their superiors or other team members about the problems, and no action was taken. Soon after, around April 3, citizens noticed a number of anomalies in the readings, including two which gave readings 1000 times higher than normal. The responsible officials handled this in an equally clumsy way, and did not make a public announcement acknowledging the problems until after the blogosphere started to freak out about the “radiation spikes.”

JB droid graph 02
Above: A graph from a JB Japan Brand realtime monitor showing an anomalous jump in readings (from a compilation at

JB droid graph 03
Above: Combined graphs from JB Japan Brand monitoring posts showing consistent anomalies. (from Twitter user 泉❀智紀(N105マスク) )

In all, 30 of the 77 new units from JB Japan Brand showed significant over-readings and other anomalies, but their data continued to be displayed on the NRC web page until Fukushima Pref. requested it to be removed on or around April 22. The NRC subsequently ordered all 77 to be removed from the web site, and after deciding that the problems could not be resolved, Fukushima Pref. cancelled the company’s contract. Company spokespeople have been reported as saying that the units were still being used in a test mode, that they were never given a chance to resolve the technical problems before the contract was cancelled, and that communication with Fukushima Prefecture had been fraught with problems and a lack of cooperation from the start.

As we discussed back in 2012, residents have been skeptical about the realtime monitoring system from the start, and this new scandal can only reinforce their mistrust of the system. It appears that bad data from 30 units was being posted on the government website for weeks. The problems three years ago with the Alpha Tsushin units stemmed from poor oversight by MEXT, which was responsible for that contract; in this case Fukushima Pref is the responsible party, and it’s extremely troubling that they did not learn from the earlier scandal and ensure that the new units were adequately tested before being put online.

Systems like these invariably have teething problems. SAFECAST has gradually been expanding its experimental deployment of realtime fixed sensors in Japan and overseas, and “living” with them for a while in order to learn their idiosyncrasies and likely failure modes. We don’t underestimate the challenges, and this newest scandal convinces us even more that citizens need to have an independent realtime sensor system they can use to check official data. At the same time we want to point out that sharp-eyed, technically knowledgeable citizens played a key role in calling attention to the problems with the JB droids. This is the kind of culture we consider ourselves a part of and want to help nurture.

Mainichi Shimbun: Glitch suspected in abnormally high readings at Fukushima radiation monitoring posts
Fukushima Minyu (Japanese)
NHK news item about the problem (Japanese)
Mainichi News (Japanese)
A compilation of Tweets in Japanese about the issue
NSR Realtime monitoring map page (old links broken)
IIDJ Japan Radiation realtime map ( data from 4500 official realtime sensors)

About the Author

Azby Brown

Azby Brown is Safecast's lead researcher and primary author of the Safecast Report. A widely published authority in the fields of design, architecture, and the environment, he has lived in Japan for over 30 years, and founded the KIT Future Design Institute in 2003. He joined Safecast in mid-2011, and frequently represents the group at international expert conferences.