SONAR AS A DEVICE FOR UNASSISTED MOBILITY OF THE BLIND

1. INTRODUCTION

Visual impairment leads to somewhat specific behaviour modes of a person. Alterations in a visual apparatus that lead to complete or partial loss of sight tend to result in more or less notable alterations at different levels of psychic and psychosocial functioning. [1] In effect, almost all spheres of a visually challenged person's life are affected.
The other source [2] claims that implications of blindness to a person's activity and individuality may be so intense in some fields, that specific discrepancy in development between the blind and the normally sighted may be debated. These discrepancies are principally noted in posture, orientation and mobility.
Orientation and mobility are essential for successful functioning of a visually challenged person, as these are essential for achieving many goals, including education, and employment.
This condition requires certain differences in education of the visually challenged persons. It is essential to broaden the aspects of person's development to self-reliance, in order to manage life successfully, and without constant assistance of the community.
Self-reliance of any person is observed through independent performance of various aspects; cultural and hygienic activities (feeding, personal hygiene, public appearance, clothing), mobility, finances, domestic affairs etc. Reduced or insufficient self-reliance is noted amongst all age groups. Most prominent factor in the visually challenged person's self-reliance is still mobility and orientation, both in known and unknown surrounding.

Fig. 1. Typical white cane.

From times immemorial there is evidence of various aids in form of sticks that were used as mobility aids for the blind. Even today the most widespread mobility aid is a white (long) cane (Fig. 1.). There are only minute differences between nowadays white canes and any stick found in nature, in form of quality or some custom adjustments.
In order for a blind person to move about safely, it is necessary that some mobility aid is used. These aids may be in forms of [2]:

- a human guide with normal sight;
- a white cane;
- some electronic device;
- a guide dog.

Each of these mobility aids has its own advantages and drawbacks that are not in the focus of this paper. Our primary goal is finding a simple and successful mode of improving orientation and mobility amongst the visually challenged population, using efficiently all human and material resources that are put to our disposal.
It is worth mentioning that there is a great segment of the blind population that commute independently, some of them even very successfully, but all of them desire to have some solution that will be safer. White cane protects only lower body parts, and head remains unprotected. One may extend his/her arm to protect a head, but then both arms are engaged. It is also important to note that the tactile range of a white cane reaches up to one step ahead, and everything else is unknown (left or right, the other side of a street, buildings, staircases, people etc.).
Some 15 per cent of the blind use guide dogs to improve their mobility, as the dog recognises all barriers on the path and avoids them. A blind person in this case loses immediate contact with the surroundings, because the dog is unable to describe the nature of obstacles and whereabouts.
Spatial perception is essential in the development of the blind, but it falls short due to the lack of aids designed for this purpose. It happens that even adults are unaware of their surroundings, even in spaces they occupy daily. This leads to riskier mobility, and injuries caused by some unaccounted for objects. Children with sight deficiencies are even more vulnerable as their spatial perception is limited to tactile sensations, linearly with auditive backup as the only source of information about their surroundings. Auditive information is exactly an approach to developing spatial perception, as this may be related to the visual information received by a normally sighted person.
Up to date, there were various approaches to the development of devices for improving mobility in form of electronic canes with various device-to-human interfaces. Information about obstacles was conveyed by means of audible signals or vibrations.

Designing of test sonar is currently in progress in Croatian Association of the Blind (CAB), with purpose to facilitate mobility of the blind, both indoors and outdoors.
The next stage of a project is testing of a device by a professional peripatologist (orientation and mobility trainer). It is assumed that this person is competent for conducting this evaluation due to his/her assisting the blind on a daily basis in developing their perceptual and motional skills by means of using different aids and techniques, all of it in order to develop a blind persons self-reliance in his/her mobility [3].
Testing is followed by training of a group of some 10 blind persons for using of a device in a period of up to 3 months, and then they will continue testing independently (circa 1 year). Tests will be performed in all given surroundings and situations, and their experiences will be recorded in a test log. It is also essential to collect experiences in using of sonar in correlation with other aids, including a guide dog, and a white cane. No illicit interactions are anticipated, but they will be cross-tested for compatibility anyway.
Our only concern regarding interactions so far was the dogs hearing frequency response, illustrated in Fig. 2., with curves representing different dogs' response [4]. It indicates that the safe range is beyond 30 kHz, which is compliant with the sonar concept.

Fig. 2. Dogs' response audiogram

With feedback from field testing, eventual flaws in the initial design will be corrected, and a new batch of perfected sonars will be produced.
The following step will be introduction of a perfected device to blind children, in order to develop their spatial perception. We feel that this stage is a paramount of this project, as it enables teachers and children a successful realisation of this extremely sensitive matter, and their further development in general. We expect all this to lead towards a better integration of the blind into society, and developing confidence towards the blind from society in general.

A peripatologist (orientation and mobility instructor) creates an individual training plan for each blind person participating in training. These plans are created upon a thorough medical and motional sensory skills check-up, which include awareness of surroundings, and personal motivation.
Instructor will also assess the choice of aid that is optimal for each individual. Upon the training conclusion, instructors remain reachable to the trainees, and in cases of real problems they will assist in finding solutions [3].
In order to yield as accurate information as possible regarding the use of sonar in the field, certain tests will be prepared by the instructor. These tests will assess the trainee's success in detecting various obstacles indoors and outdoors, and in different heights (various road signs, open windows, parked cars, trash containers, walls, and different objects lying on the floor), or velocities, including vehicles, humans or animals. Tests will establish optimal sizes of immovable objects useful for orientation in space, their significance and relevance.
It is important to evaluate usefulness of sonar in detection of obstacles in form of stairs, and glass. Optimum volume of received tone must be established, in order not to mask environmental sounds - essential in orientation.
Atmospheric influences may affect sonar's operation, due to its electronic parts. Functioning in harsh environments and resilience to shocks must also be assessed.
From all these assumptions it is most important to establish sonar's usefulness in different environments, and its optimum purpose, as opposed to other aids in their optimum areas of use. It must be clarified that sonar is not intended to replace white cane, but to supplement it in areas of orientation, collision avoidance, and obstacle recognition. Recognition of passing by humans will certainly aid to the blind person's social skills, and acceptance by the society.

Successful sonar designs date back to 1965 with a patented design by Kay L. et al. [5]. Whole development up to nowadays was based upon this original design, with advances mainly in the field of improvement of electronic components or some assemblies. As a seminal work, only this reference is quoted.
Leading idea in perfecting of our design is extending its availability beyond economical constraints. Only of-the-shelf components are used throughout design. In order to make it even easier to reproduce, no programmable devices are used.
The principle is often regarded to as Frequency Modulated Continuous Wave (FMCW), modulation is a falling frequency sweep, and two separate ultrasonic transducers are employed for a full return path. Both receiving and transmitting transducers are standard ubiquitous 40 kHz piezoelectric transducers, as in reference [6], chosen for their availability, safety and low price. The rest of electronics is designed to accommodate for this choice.
Narrow bandwidth associated with these transducers is compensated by de-tuning the transducers, and by signal strength equalisation. Somewhat better transducers are available, but with expense of additional safety measures related to bias voltages exceeding 100 V, and therefore these were discarded as unwelcome.
This family of sonar designs employs human hearing ability to discern several sound sources separated by frequency. Different objects in space reflect ultrasonic wave with different delays, and from different angles. Due to the directivity of transducers set-up, there is a possibility to distinguish objects spatially if reception is binaural. Received sound pressure is also proportional to each object's reflectivity, and pitch is directly proportional to distance.

4.1. Method description

Falling frequency sweep is transmitted towards the object that reflects it partly towards the receiving transducer. Originating sound is received together with the delayed reflected signal, and in effect they form an AM composite with falling sweep frequency as a carrier. Simple AM envelope demodulation produces audible response of the environment. As the frequency range is limited, pauses are inevitable.

Fig. 3. Frequency-time plot.

Sweeping range and repetition frequency are adjusted to accommodate audible response to the observing range. With human hearing range in which sound-emitting objects may be spatially discerned extends roughly from 400 Hz to 4 kHz, there is one decade of distance covered by one repetition frequency adjustment. It is anticipated that only two ranges (e.g. 0.3 m to 3m and 1 m to 10 m) will suffice.
Falling sweep is chosen for Doppler shift reasons (Fig. 3.). With this choice, Doppler shift of approaching objects is increasing the stationary pitch. If the choice was raising sweep, Doppler shift of approaching objects could lower the stationary pitch down to inaudible frequencies, making this device prone to causing unintended collisions.
With low speed of moving target (vt<(1)
Delayed wave's frequency is higher than the carrier due to the falling sweep, and thus the approaching objects will have higher resulting pitch than departing ones at the same distance.
Most bats use exactly the same sonar principle as described here, all of them employing the same falling sweep as described before. As all the analogies apply, their behaviour will be used as a model to illustrate aspects of sonar use.
Bats usually reside in caves in large colonies, and yet they find their way in and out easily. Therefore, it is not anticipated that interference with other sonars, even in confined areas will be a problem. When testing a prototype, some sounds related to bats were observed, but completely different than surroundings response.
When hunting, bats are aware of their prey species by means of Doppler shift of buzzing insects' wings. Similarly, humans tend to wave hands when walking, giving a walking pedestrian unique signature.
Stationary objects posses signatures related to their texture or structure. Rough surfaces response is noisy, with grass and trees being pure noise. Stairs and fences with repeated structure sound like organ, and isolated poles sound like bells. All Doppler related phenomena are superimposed to these, making them clearly distinguishable as related to motion.
A step forward in this project is design of acoustical catadioptric reflectors in a form of small and dispensable objects, used for exaggeration of acoustic signature at some important places e.g. traffic poles at crossroads, important corners, building entrances etc. These reflectors will be designed to withstand weather and mechanical deformations, with concern to economical issues and ease of application.
Sonar device is going to be attached to a baseball cap, and batteries with control electronics will be placed into a belt bag. This choice is aesthetically tolerable nowadays, and feedback from users on this solution is welcome. Choice of batteries is going to be limited only to their supply voltage, in order to make use of their wide availability. A blind user will be supplied with all spare assemblies, in order to increase his/her service autonomy. This is estimated as far cheaper solution than producing a device with comparably better reliability.

4.2. Economical concerns

All economical concerns are born in mind in order to protect the most vulnerable part of population - children. Luxurious designs will find some market, but tend to be extremely expensive, as the total market share is extremely small. Total population of the blind is about 0.1 to 0.15 per cent, depending on the criterion of blindness. Children are even smaller fraction of this number, and their financial abilities are limited to the good will of society, and resources of their parents.
Making this project an open source will eventually keep prices down, and development constant. Making sonar for a blind colleague could become a rewarding DIY project, for the benefit of both.

As the sonar device is still under construction, its total potential is still unknown. It is still uncertain if the visually challenged person will accept it and use it, and what will be the considerations of its use, training of trainers, training theatres, or financing. It is a potential of this new approach that we explore and results that are gathered on its way that will lead to progress.

[1] V. Stančić: Oštećenja vida - biopsihosocijalni aspekti, Školska. knjiga, Zagreb.. (1991).
[2] G. Zovko: Peripatologija I. Školske novine, Zagreb. (1994)
[3] V. Mršić: Orijentacija i mobilitet u Hrvatskoj. Obučavanje slijepih i slabovidnih za neovisno kretanje. Hrvatska udruga za školovanje pasa vodića i mobilitet, Zagreb, (1995).
[4] http://www.lsu.edu/deafness/HearingRange.html
[5] L. Kay et al: "Apparatus for furnishing information as to positioning of objects" US Pat. No. 3,172,075, (1965)
[6] http://www.massa.com/datasheets/tr89.html