A smart landslide detection system developed by a research team led by Professor Cheng Chun-hung, Associate Professor of the Department of Systems Engineering and Engineering Management was recognised as one of the latest application-oriented researches that ‘save citizens’ lives’ and has recently won a silver award under the category of Smart Living (Smart Home) of the HKICT Awards 2018. 

Mountainous terrain covers 60% of the land area of Hong Kong, and landslides may happen under conditions of continuous rainfall. If early signs of debris flow are not detected, it might cause casualties and losses. Professor Cheng Chun-hung and his team have successfully developed a smart landslide detection system to monitor real-time terrain conditions and transmit wireless data to the mobile devices of duty staff, using low-cost and low-energy sensing devices. Last year, in collaboration with the Geotechnical Engineering Office under the Civil Engineering and Development Department (CEDD), CUHK conducted experiments in Sham Tseng where landslides had occurred. The system performed steadily under bad weather conditions during the experiment, demonstrating its ability to immediately notice any landslide activity on hillside near residential areas. 

CEDD has implemented risk mitigation works for 30 natural hillside catchments per year since 2010, and thus far there have been over 200 debris-resisting barriers well built in the forms of rigid concrete barriers and flexible wire-mesh fence barriers. However, the diverse and remote locations of the different barriers and the variations in weather conditions impose practical difficulties for performing onsite checks for any accumulation of debris from natural landscape erosion. 

In view of this, CEDD joined hands with CUHK Department of Systems Engineering and Engineering Management to design a landslide detection system mounted on debris-resistant barriers to monitor and give notice of any landslide activity. The system was piloted on the hillsides of Sham Tseng in November 2016. The pilot system, with sensing devices and system components, has been installed on 4 out of 19 debris-resistant barriers there. The sensing devices are designed to detect any impact when landslide debris hits the barrier wall, measure the debris accumulation level inside the barrier, and detect any inclination of the wire-mesh fence. In addition, there are infrared night vision cameras to provide a 24-hour visual context for landslide situation assessment. The authorities concerned can then be alerted at all times to any landslide activities in remote hillsides via the Landslide Detection System (LDS) mobile app in both iOS and Android platforms.  Most importantly, it helps proactively provide timely data, especially under extreme weather conditions, to assess any landslide situation close to residential areas. 

The hilly terrain and large spacing have made generating the power supply and communicating with wireless networks the two biggest challenges to developing the landslide detection system.  On one hand, the team adopts low-voltage sensors, camera electronic components, and low-power consumption electronic components that are commonly available, low-cost and energy-efficient.  The battery supply for the system can last for three years, which, in turn, saves much maintenance and consumption cost.  On the other hand, the team wisely explores the use of solar panels for a steady electricity storage and supply.  Under normal circumstances, the system can continuously operate for 15 days, with sufficient solar supply. 

Regarding the dispersed locations of barriers and other connection problem of the smarts system, the team turns to LoRa which is a reliable and low-energy wireless technology. The wireless communication has proved successful in transmitting complete data back to the base station in extreme weather conditions and hilly environments. The data received can also be sent via mobile to staff on duty. 

Dr. Dorbin Ng says that the landslide detection system is not only flexible for deployment, but also very reasonable in terms of cost and energy consumption level.  The communication and sensing performance during the typhoon signal number 8, or higher, remains satisfactory.  The pilot scope has recently been extended to the hillside at Queen Mary Hospital in Island South.  He hopes that the pilot system at the new barrier built on the hillside behind the hospital can provide more data for future large-scale deployment.

A research team led by Prof. Raymond W. Yeung, Choh-Ming Li Professor of Information Engineering and Co-Director of the Institute of Network Coding (INC) at CUHK, has invented BATched Sparse code (BATS code) to improve the network transmission rate of networks with packet loss. It is one of the world’s most mature networking techniques in which transmitted data is encoded and decoded to increase network throughput, reduce delays and make the network more robust. The potential application of BATS code includes a wide spectrum of areas such as Internet of Things (IoT), fog computing, underwater and satellite communication systems and disaster communications, breaking geographical boundaries with technologies. The team is now in the process of applying to Institute of Electrical and Electronics Engineers (IEEE) to make BATS code one of the international wireless communication standards. 

Pioneering Network Coding Technologies 

As an efficient linear network coding solution, BATS code replaces routers with coders that transmit ‘evidence’ of a message instead of sending the message itself, and overcomes the problem of data loss during wireless transmission, and offers higher speed, reliability and stability. Compared with conventional random linear network coding, BATS code offers a lower encoding and decoding complexity, and requires a much smaller buffer size at the intermediate nodes. For example, for a multihop network, BATS code can increase the transmission rate by 56% and reduce the loss rate by 29%. With 7 years of Prof. Yeung’s team effort, the performance of BATS code has proven remarkable in making network communications more efficient, reliable, stable and secure. 

BATS code not only helps to compensate for the network loss, it improves the throughput for multicast with just minimal storage capacities and hardware requirements. In principle, a basic low-power communication hardware will suffice to consistently achieve a high transmission rate for multi-hop wireless networks with hundreds to a thousand nodes (the number of intermediate nodes or liner networks). In other words, this makes BATS code ideal for implementation on multi-functional smart lampposts especially when extra large-scale infrastructure and fiber network upgrade are not needed for transferring data to the server. More importantly, the advantages of BATS code, which include high throughput, low latency and low storage requirement, strongly favour transforming Hong Kong into a smart city. 

Cross-disciplinary Research Collaborations

BATS has already obtained a number of patents in Europe and the US. Thus far, the team has been actively in touch with the Hong Kong Applied Science and Technology Research Institute to explore the opportunities for collaboration on their smart city’s lampposts scheme, with Beijing’s aviation technology units for its application in satellite communication, and with the University in Nova Scotia, Canada for testing of acoustic signals in underground communication. BATS code can particularly tackle the problem of data loss in the ocean, and, in turn, facilitate the civil monitoring process on underwater builds. Looking ahead, Prof. Yeung hopes that by introducing BATS code into hot air balloons or aerial photography, setting up wireless networks and emergency rescue and communication works will become less difficult in remote and hilly districts, and that BATS code will have great potential for benefiting mankind.

Meanwhile, the team has been invited to work in different fields on promoting fog computing, a conceptual extension of cloud computing. The objective is to reduce the network congestion and streamline the algorithms and layering. CUHK and INC are the only two institutional members of OpenFog Consortium in Hong Kong. Prof. Yeung remarked, “There is a strong correlation between fog computing and Internet of Things (IoT); for example, the industrial and medical application of smart robots and Internet of Vehicles (IoV), which came under the umbrella of IoT. BATS code is one of the keys to making IoT possible”. 

The network coding theory which originated at CUHK is a major breakthrough in information sciences. Its fundamental concept was introduced in the late 1990s, largely due to the work of Prof. Raymond Yeung, and has become an innovative network coding technique to revolutionise the world of wireless communications. Prof. Yeung hopes that with this breakthrough technology from CUHK, BATS code will be shared to realise Hong Kong’s smart city vision and its infrastructure-based development, and to make Hong Kong a world-class smart city.

Demonstration of a video streaming comparison of BATS and fountain codes in a two-hop network can be viewed at http://iest2.ie.cuhk.edu.hk/~whyeung/BATS.mp4