WHETHER it’s the mountains of Munnar in Idduki district of Kerala in the south or the Himalayan ranges in Uttarkashi region in the north, landslides are a common occurrence. They cause mass destruction year after year. The problem is that landslides are difficult to predict as they are very localised and depend on the soil conditions prevalent in a region. The methods in use to predict landslides depend on satellite data and geographical information system. But the results are often inaccurate as the methods study conditions superficially and fail to analyse what is happening at the micro-level.
To help predict landslides, scientists at Amrita Centre for Wireless Networks and Applications, a part of Amrita University in Kerala, have devised a system based on wireless sensor network. It is a first-of-its-kind system based on wireless technology in India and among the first in the world for predicting such mishaps. A wireless sensor network uses a variety of sensors to monitor physical conditions in an area. They can sense the changes in soil pressure, moisture, movement and transmit the information to a central processor. They are widely used for environmental monitoring, such as detecting levels of pollutants and forest fires.
How it works
The prediction system has been installed at Munnar’s Anthoniar Colony and consists of 150 geophysical sensors. These include moisture sensors which determine soil moisture by measuring dielectric constant (an electrical property dependent on moisture), strain gauges that measure strain in the soil, piezometers which measure the water pressure in soil pores, tiltmeters, that measure changes in soil-level, rain gauges and geophones, a device which converts landmass movement into voltage.
The sensors are connected to form 20 wireless sensor nodes. Each node is a combination of a sensor, a transmitter, an electronic circuit for interfacing with the sensors and an energy source, usually a battery. The real-time data generated by these sensors is sent to a field data management centre located a kilometre away and then transmitted to the University campus, situated about 252 kilometres away, for processing.
The university has developed a computer model to assess the data and issue warnings. For example, when the level of moisture increases over a certain level, the system raises an alarm.
The first-level warning is issued when the soil has more than 90 per cent moisture. Second-level warning is issued when the pore pressure value (pore pressure increases with rainfall) increases. This way, a warning can be sent out to the district authorities around 24 hours before a landslide occurs, giving enough time to evacuate people.
Costly affair
V S Prakash, director at Karnataka State Natural Disaster Monitoring Centre, points out that no compromises should be made when it comes to protecting human life. But the system is expensive, he says. It costs around Rs 5 crore. The researchers are trying to reduce the price by replacing costly sensors with cheaper alternatives. For example, the inclinometer used in the system costs around Rs 1 lakh. “It has now been replaced with tiltmeters which cost Rs 25,000 and provide equally good information,” says Maneesha Ramesh, director at the Centre who has set up the system at Munnar. She hopes the system can be replicated in other landslideprone areas. “The data produced by the sensors can be used in other fields like agriculture, where knowing the physical properties of the soil is crucial,” she adds.
M H Mehta, chairperson of the National Bioshield Society who is looking for ways to reduce the harm by landslides in Uttarkashi, says, “The experiment provides an opportunity to take steps to prevent landslides.” He suggests more trees should be planted to improve soil binding.
Other factors
Cost is only one of the concerns. Before such a system is used in the field, it is important to ensure that the data being transmitted by the sensors is secure, says Dawn Song, an expert in software security at University of California, Berkeley in the US. For this, one can incorporate protective software like BitBlaze that analyses malicious software code and WebBlaze that focuses on defending web-based applications and services against malicious software.
During a disaster, there is also the problem of overflow of data and lack of resources to handle it. Though ad hoc networks are generally put in place, these have limited bandwidth and crucial data can get lost. To avoid this problem Faisal Luqman and his team at Carnegie Mellon University in the US have developed a framework that is loaded on a smart device and prioritises data. For example, priority is given to information from a sender with low battery levels.
Apart from being used in disaster situations, such wireless technologies can also be put to use in an urban set up for managing traffic for instance, says P Venkat Rangan, chancellor of the university.