Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
2l1 TELECOMMUNICATIONS RESEARCH STUDY I. General Remarks After reviewing the summaries of telecommunications research and development provided by the Department of Defense (DoD) and NASA, one is impressed by the sheer magnitude of the effort being sponsored by these agencies. When one adds to this the extensive basic and applied research effort at Bell Laboratories and the research and development programs of the independent laboratories, it would appear that telecommunica- tions research in the United States is being carried on at a very substantial level. It is true that DoD and NASA sponsorship is (and probably should be) directed increasingly toward mission-oriented applications, but it must be acknowledged that in the past much of the best basic research has been sponsored by these agencies. Supposedly, it was the intention that the reduction of basic research support by DoD and NASA would be made up by increased allocations to NSF, but in fact the increase to NSF has been but a small fraction of the amount required. The net result is that the support for basic research in this country has decreased drastically. Undoubtedly the panel will wish to address this problem in formulating its recommendations regarding telecommunications research. How to do this without appearing to come up with "just another self-serving committee report" will be a challenge which must be met. II. Suggestions for Future Research The following are some suggestions for research directions , or specific research areas suitable for work in a University laboratory. It must be acknowledged that the present University research effort in telecommunications is miniscule compared with that of an industrial laboratory such as Bell Laboratories, but for reasons indicated elsewhere in this report, it is important nonetheless.
2l2 A. Message Processing for Efficient Transmission A great portion of messages transmitted are highly redundant. This is especially true for speech and image signals. Reducing this redundancy reduces the required capacity of the communica- tions system many fold. Research is strongly needed in fast and real-time algorithm which can be implemented on a small computer. B. Data Traffic Organization and Message Switching Transmission facilities can be minimized by sharing. Traffic flow in a network can be organized to increase thruput and reliability. Methods of optimum control of traffic flow in a network are urgently needed. Some consideration should be given to dynamic relocation of transmission paths. C. Network Optimalization Sharing of transmission facilities can be accomplished by networks of channels which share the flow of messages. Research is needed in the optimalization of design layout which increases message flow and makes optimum use of transmission facilities. D. Transmission Reliability, Coding and Security In data transmission reliability is extremely important. We need methods of ensuring accuracy of transmission, including error detection and retransmission and forward error correction. Emphasis should be on simple methods that can be implemented easily and economically on a computer. E. Efficient Modulation Methods Efficient modulation methods should be developed to make the best use of frequency bandwidth. Multiple level schemes should be developed to take advantage of the high signal to noise ratio of some transmission media. F. Pulse Code Modulation Time domain multiplex methods should be studied along with the synchronization of such systems. Generalization of Pierce's
213 model and Sandberg's model should be investigated in a manner that includes practical sources of error. G. Digital Filtering A better understanding of the errors in digital filters is needed. This aspect is crucial in determining the per- formance of digital filters. H. Electromagnetic Environment Control In order to control the quality of the electromagnetic environment the best approach seems to be the use of adaptive power control, i.e., reducing the radiation of each trans- mitter to the minimum required level of radiation which will achieve the desired quality of received message. The appro- priate research effort is, therefore, to seek effective algorithms and system designs for accomplishing this type of control and to analyze the dynamic performance of such a system. I. Low-loss Microwave and Millimeter Wave Transmission Of the many papers presented at the l972 International Scientific Radio Union (URSI) General Assembly in Warsaw, two hold out prospects for practical transmission systems with reduced attenuation at millimeter wave and microwave frequen- cies. The "frame beam-waveguides" described by an Italian group from C.N.R., Florence, P. F. Checcacci et al, represent a new class of the beam waveguide, originally described by Goubau. The new class consists of a series of equispaced dielectric frames only slightly larger than the beam cross section. Experimental tests show a measured attenuation about one-half of that for an iris waveguide having the same cell sizes. A new type of coaxial waveguide described by H. M. Barlow of University College, London, is capable theoretically of sustaining "dipole-family" modes,in addition to the usual circularly-symmetrical waves and the waveguide modes that arise when transverse resonance is permitted. The lowest order screened dipole mode propagates (theoretically) throughout the spectrum with an attenuation of about one-third or less of the corresponding circularly symmetrical mode. The practical problem that remains to be solved is the construction of coaxial
214 guides whose walls have the required (non-zero) reactive im- pedence demanded by theory. Although skepticism is likely to prevail until the theoretical results have been verified experimentally, the goal of reduced attenuation is of suffi- cient economic importance to warrant considerable further effort. J. Telecommunications Education The need for more engineers trained specifically in telecommunications has been raised by Dr. Kandoian and other members of the Panel. There is good reason to believe that an increased output of telecommunications engineers from some of the larger departments of Electrical Engineering could be achieved fairly easily and quickly. Most departments have a flexible curriculum which permits a considerable degree of specialization in the senior year. In addition, a year of graduate study in telecommunications engineering could provide well-trained experts at the M.S. level. What is needed, besides faculty competence and interest (both of which exist at many schools), is modest support for research projects and graduate assistants in this area. in most of the better schools all new staff are expected to engage in research as well as teaching in order to maintain knowledge and competence in their fields. Faculty salaries can be paid largely from University funds, but support for research equipment and research assistants must come from Federal or industrial sources. Funding of modest telecommunications research grants by NSF, for example, could have a marked beneficial effect on the output of telecommunications engineers. K. Devices for Telecommunications For an authoritative, comprehensive survey of the needs and problems in the device area, one can hardly do better than to examine the appropriate sections of a study by a distinguished panel of experts under the chairmanship of the late jack Morton of Bell Telephone Laboratories. This report of the National Materials Advisory Board entitled "Materials and Processes for Electron Devices" was published in May l972 by the NAS-NAE. It surveys in depth the materials and processes for solid state electronic devices, identifying problems, considering possible solutions, and making recommendations. Edward C. Jordan Department of Electrical Engineering University of Illinois, Urbana
2l5 COMMENTS BY E. RECHTIN ON "TELECOMMUNICATIONS RESEARCH STUDY" Message Processing for Efficient Transmission A better case can be made for the high redundancy of speech than for that of images, unless the latter are either stationary or move very slowly compared to the picture framing rate. Some speech compression systems have been built which show that essentially the same quality of speech can be achieved at 2400 bits per second as at (uncoded) 48,000 bits per second; a further compres- sion of more than a factor of about l0 seems unlikely for any reasonable cost. Thus, for speech, the problem is to realize this factor of a l00 or so in a practical and inexpensive way. Efforts to compress typical television signals show less promise; the exception is fixed dis- plays and computer outputs for which the reader can afford to wait a few seconds. Data Traffic Organization and Message Switching Another technique of interest is "saturation routing" in which the telecommunication system is probed for open pathways between signal source and the signal recipient. The algorithms for this dynamic relocation of transmission paths in a large communication network do require research. Network Optimalization Another challenge is the design of "precedence sys- tems" in which higher priority calls can preempt those of lower priority. Existing precedence systems are relatively primitive and tend to saturate during crises when all the users escalate their precedence until their calls get through. The call completion rate for most users,therefore, drops abruptly during these conditions. Transmission Reliability, Coding and Security Some of the better and more recent payoffs have been in the treatment of the kinds of non-Gaussian errors which are characteristic of many channels, including telephone circuits, HF links, and ELF links. These links are characterized by "hits" rather than a typical Gaussian "noise level" with results that are particularly bad for digital transmission.
216 Efficient Modulation Methods We have tended to concentrate on optimal use of an assigned frequency band rather than on minimizing the channel to channel interference. It might be an inter- esting research project to derive the best in-channel modulation, subject to the constraint of specified out- of-channel interference. Pulse Code Modulation Of particular importance these days is the syn- chronization of pulse code modulation systems for mobile platforms, particularly aircraft. E. Rechtin Department of Defense (Telecommunications)
