Reversing a long-term downward trend, crime rates are slowly beginning to increase in some areas in North America. With local, city, and state government budgets tight as ever, law enforcement and public safety agencies have to find new cost-effective strategies and tactics to deter and solve crime. Given the growing limitations on manpower, our officers cannot be everywhere at once.
One method that law enforcement is increasingly deploying to combat crime is video surveillance over broadband networks, which are increasingly deployed as part of IP-based public safety radio networks.
The increased use of IP technology in public safety networks as well as other information technology offerings has provided first responders with the opportunity to deploy surveillance capabilities that were not easily implemented with closed circuit television (CCTV), which is transmitted over traditional analog frequencies. The adoption of IP and wireless technology has expanded the options for transporting video, and its digital nature has opened the path for intelligent analysis of the content. The net outcome is that video surveillance can be deployed in more places and be monitored more effectively, with fewer people staring at a wall of monitors. Ultimately, video surveillance has the potential to become an immediate response tool, becoming a killer application that aids crime prevention and criminal apprehension, rather than the current practice of using video to identify the “ones that got away.” How Does it Work?
A typical IP video network includes a collection of fixed and pan-tilt-zoom (PTZ) cameras (either IP cameras or analog cameras attached to IP encoders), network video recorders (NVR) and viewing platforms. The actual hardware may differ, depending on the vendor or combination of vendors, but across all the installations is an IP network, which actually transmits the video. While the video traffic is usually shared with other network traffic such as data and voice, the bandwidth required for video has a tremendous impact on network design and performance. Wireless broadband links have added a new level of cost-effective flexibility to video surveillance. The challenge for the network, however, is to maintain acceptable performance. http://www.macom-wireless.comWireless Video Applications
On popular television shows and movies, both law enforcement officers and criminals can quickly set up a system that remotely transports video almost anywhere. In many respects, this scenario exceeds technical reality. While covert video is prevalent, wireless video is challenging, and vehicle systems today almost always rely on local recording. But recent innovations in broadband wireless communications do have the potential to bring these applications closer to Hollywood’s promise of live, streaming video surveillance. Surveillance
Surveillance is the monitoring of remote locations in order to send visual information from one place to another. Its uses range from monitoring traffic in cities or highways to monitoring people in almost any location. Locations are chosen due to their history of criminal activities or their vulnerability to terrorist attacks. In addition to the obvious security applications, video surveillance technology is used to monitor traffic flow and pedestrian congestion in public spaces, to patrolling borders and facility perimeters and provide secure perimeters around public events. Cameras deployed for these purposes may be fixed and set to view the specific region of interest or PTZ to allow redirection of view. There are even “gunshot detectors” that attempt to recognize and direct cameras to the sound of a discharging firearm. Facial-recognition programs attempt to pick a face out of the crowd with more promise than success, although under controlled circumstances, they can recognize faces. As public safety’s surveillance needs expand, so too does the number of places where a camera can be a tremendously useful tool in detecting and preventing illegal activity. However, the lack of an IP-based wireless infrastructure has hindered wider adoption of surveillance cameras. Vehicle-Based
Popular television police chase programs have effectively demonstrated the capabilities of vehicle-mounted cameras. Because of these new tools, police are also capable of sending the video, if it is of sufficient resolution, directly to a license plate recognition system, which automatically runs license plate checks to look for stolen vehicles, outstanding warrants, etc.
Buses, subways and trains have two primary video applications. One is to monitor the interior of vehicles, and the second is to monitor station platforms as passengers wait to embark and disembark. An example of this method is the well-known London subway incident in 2005 that involved the use of video to identify the suspect, albeit after the fact. Tactical Video
There are many scenarios where quick deployment on a temporary basis is critical. Hostage situations, political conventions or even public gatherings such as parades are ideal for the use of wireless video surveillance. These scenarios do not usually allow the luxury of establishing wired connections, and due to their immediate nature, they require quick and reliable deployment in order to maximize effectiveness. Analytics
Many commercial and law enforcement applications require automated surveillance systems. Mounted video cameras do not incur large recurring costs; however, using human resources to observe the video does. Although surveillance cameras are already prevalent in banks, stores, and parking lots, video data is usually used “after the fact” as a forensic tool, and it loses any active, real-time benefit. The goal of continuous 24-hour surveillance is to alert law enforcement and security officers to a burglary in progress or to a suspicious person loitering in a parking lot while there is still time to prevent the crime. Intelligent algorithms can analyze video content and trigger alarms based on suspicious activity. Typical analytics include motion detection, missing or abandoned objects, boundary crossing and counting events. Embedding these real-time analytical algorithms into video encoder allows alarms to be triggered immediately and automatically, initiating preventative, proactive police action. Video Network Traffic
The basic parameters of network engineering, bandwidth, latency and jitter, need to be considered when crafting the technical requirements to transport video over an IP-based network. Even with the advanced codecs (compression / decompression algorithm) that are used today, video intrinsically consumes a large amount of bandwidth compared to other typical network applications. Adding a number of video streams without proper planning can quickly cause network congestion, even on a robust, wired network. This issue is obviously more critical over wireless networks, which have less flexibility and capacity. The good news is that advanced codecs can bring bandwidth usage down to manageable levels, and once the wireless segments are properly engineered, the rest of the network usually will not have a traffic-loading issue.
The bandwidth required varies widely based on the resolution, frames per second (fps) desired and codec used, as well as the motion content of the video. The motion content is normally characterized as low, average or high, and those classifications are very subjective. The level of compression varies with codecs. General bandwidth requirements for various applications of a well-implemented MPEG4 and are shown in Table 1 (IP protocol overhead not included). Note that the particular implementation of each vender determines how the codec behaves as the video image needs more bandwidth than has been set. Artifacts such as jerkiness, blocking and pixilation may be introduced. The better codecs have higher initial compression and introduce less objectionable artifacts when the bandwidth constraints are reached, and often allow selection of which parameter to squeeze first. Resolution
SIF (Standard Interchange Format or Source Input Format) or CIF (Common Interchange Format) is a measure of video resolution. The two are virtually identical and are used interchangeably. SIF resolution measures 352 x 288 pixels for PAL cameras or 352 x 240 for NTSC cameras.IP Video Codecs and Compression
A video codec is designed to compress and decompress digital video in order to reduce the amount of bandwidth required to transmit and store the video. This is needed as the raw data rate of uncompressed active digital video can be in excess of 160Mbps. MPEG
The Motion Picture Expert Group (MPEG) generates standards for digital video and audio compression. There are several MPEG codecs, with MPEG-4 being a very popular high-compression video codec standard. There are different implementations, profiles and options that are somewhat compatible. The standard does NOT dictate how the video stream is placed into packets, nor any of the watermarking and digital signatures applied to prevent evidence tampering; so in the public safety video world, open interoperability is between cooperating vendors. The latest codec, called Part 10, is the same as H.264, and provides a 20% to 40% bandwidth improvement over the MPEG-4 Simple or Advanced Simple Profiles. Each generation of codec uses improved compression techniques, while building on the prior generation’s standards.
Some of these internal codec details include:
• MJPEG—Motion JPEG where each image is compressed with JPEG and all images are transmitted;
• And MPEG-2—standard digital cable, DVD or DTV, which uses I frames (full image frame), P frames (predictive frame from prior frame) and B frames (interpolated from both prior and following frame). Each frame is processed as 8x8 blocks. Typical compression results using MPEG-2 are:
• Cable—352x240, about 4 Mbps;
• DVD—352x240, 352x420, 740x468, up to10 Mbps;
• MPEG 4—incorporates frequency domain prediction with a finer resolution of picture elements and motion vectors;
• And MPEG-4 10, H.264—includes spatial domain prediction, variable block size and a host of other enhancements.
The Wireless Network
There are a number of tools available to design a wireless network to reliably carry video traffic. The first is video codec implementation, where the settings provided for frame rate, resolution and target bit rate allow trade-offs to be made with video quality and the traffic applied to the network. The second results from the QoS (Quality of Service) provided by the WiMAX (Worldwide Interoperability for Microwave Access) protocol, which is utilized in Tyco Electronics’ M/A-COM VIDA (Voice, Interoperable, Data and Access) Broadband network. The bandwidth available via each wireless link can be dynamically allocated to the multiple traffic flows according to the QoS parameters configured for each. Thus, the capacity is assigned according to pre-configured priorities, rather than simply channel contention or congestion effects.
Different applications have different requirements regarding handling their traffic on the network, and QoS determines the ability of the network to meet a given traffic service level requirement in terms of throughput, latency and jitter. Requirements are defined in terms of these QoS parameters:
• Latency: How much time it takes an IP packet to be transmitted from Point A to Point B;
• Jitter: The variation in latency among packets;
• And bandwidth: The rate at which application traffic is carried by the network.
Guaranteed QoS is assured as all communication scheduled by the base station(s), is synched with contention slots that are provided for subscriber stations (clients) re: the requested bandwidth. This coordinated scheduling feature of the WiMAX protocol provides significant advantages by:
• Eliminating contention between registered clients;
• Maximizing channel utilization;
• Enabling guaranteed bandwidth services for critical multi-media applications; • And offering four QoS classes that enable customization for specific mission-critical applications.
802.16 WiMAX vs. Other Protocols
The foregoing discussion mentions some of the WiMAX capabilities that contrast to Wi-Fi. Learning from the Wi-Fi experience, WiMAX purposely includes strong security tools, such as X.509 digital certificate authentication and encryption. WiMAX is a scheduled protocol, so the contention DoS (Denial of Service) and heavy traffic issues are avoided, and true QoS can be guaranteed. These are tremendous benefits for wireless video. The drawback is that the network has to be designed with this functionality in mind. Additionally, the desired service flows must also be configured. As a result, the ad-hoc network formation is not as free form as Wi-Fi or mesh protocols. In reality, this is a minor point, as the ad-hoc feature only works well when the wireless network is very lightly loaded. The 1/n multihop capacity degradation along with the contention issues of Wi-Fi and mesh topologies force some upfront design and adding WAN backhaul points to reduce the number of hops and accommodate the traffic load. The bottom line is that the spectrum is not used as efficiently, and the ad-hoc benefit is counter to providing user QoS, which makes video possible.
Advanced video compression techniques such as MPEG4 and H.264 have dramatically reduced the bandwidth requirements for video, to the point that effective transport and delivery of remote video over wireless links are now a reality. Video traffic, however, is still a significant drag on the flow of network traffic and is still subject to real-world bandwidth limitations and propagation constraints. Thus, link impairments and bandwidth contention must be accounted for in the system design of the network. This trade-off encompasses RF path engineering, network topology and traffic engineering, QoS mechanisms and IP networking.
Comparison of 802.11 protocol-based networks and 802.16 protocol-based networks shows that while the 802.11 networks can transport video as an incidental “non-critical” service, they lack the intrinsic security and true QoS to provide reliable operation so important to public safety. The QoS service flows provided by the 802.16 protocol (WiMAX) are designed specifically to address this latter issue. Security is built in, with x.509 digital certificates for authentication and AES encryption.
In the end, it is necessary to understand both the configurable parameters the video applications provide, as well as the network QoS controls to ensure satisfactory operation. A wireless network providing true QoS, such as M/A-COM VIDA Broadband, brings the reality of public safety video one step closer to the image most often represented in Hollywood.
Thomas Brown is chief engineer, Network Products, M/A-COM, Tyco Electronics.