|
Resources > Training
Shaping the Bark Behavior
How do I begin to mark the bark indication?
The bark indication is shaped by using a marker as a communication tool to tell
the animal the correct response the handler wants. The marker traditionally used
for animals has been the clicker, which is a plastic children’s toy that makes a
noise. Whistles have been used to mark behaviors with dolphins due to the
distance the animal works away from the trainer and the fact that it is easier
to hear under water. When a trainer marks a behavior the animal does correctly,
the trainer “bridges” a gap from the time of response to the time when the
reward is given. The importance of this bridge is that the animal is able to
acknowledge the correct response that produced the reward from the trainer.
To begin training the animal to acknowledge the clicker as a secondary
reinforcer, the “click” is paired with a primary reinforcer such as food, or the
toy reward. Food is easy to give and does not stop the flow of training whereas
a toy would as the trainer must stop to play with the animal after each trial.
To fire up the clicker simply do a rapid click/reward almost simultaneously
without asking the animal to do anything. The animal begins to recognize that
food will come each time it hears the click. This is conditioning the response
to the clicker by taking a previously neutral stimulus (the click) and
developing the noise of the click to function as a reinforcer after being paired
a number of times with an established reinforcer (food).
Timing of the clicker is one of the most important points of this type of
training, so exercises that will develop the trainer’s skills to deliver a quick
click at the crucial moment in training are the following:
- Throw a set of car keys up in the air and click when you catch them in your hand. Try to get this to the point where the catching of the keys and the click occur almost simultaneously.
- Drop the keys to the floor and when you anticipate the keys hitting the floor, click at the same time. You will be able to acknowledge your accuracy by the sound of the keys hitting the floor as compared to timing with your click.
Continue developing your skills until you are very accurate with clicking before training an animal. There are too many things to watch with an animal’s behavior
during a training session. New trainers frequently find themselves clicking the wrong behavior simply because their marking skills are too slow. This is a
mechanical skill that needs to be mastered in order to achieve a high level of success in training fine precision behaviors in an animal.
Your Reward System
The next important step is identifying the reward system for the animal. Each dog is different in its
desires and food tastes but most dogs will become more focused on a food reward
if they are hungry. Decrease the dog’s meals and supplement with treats during
training or use the meal times as a training session. Remember that overweight
dogs are not going to be as focused on a food reward so keep them trim. It is
healthier for a working dog to maintain a slim body and the lowered body weight
prolongs their lives and prevents the usual diseases of obesity. Divide their
daily ration into several portions to use in training sessions throughout the
day if possible. Mix up the kibble with other good treats to give variety. Types
of food rewards need to be noted. Treats can be made using chicken breasts,
steak, liver, or any meat. Sprinkle the meat with garlic and grill, then cut in
small pieces. Try to use soft food so the dog swallows easily and does not choke
when trying to eat and bark. Other ideas are hot dogs, cheese pieces or cheese
whiz, baby food, tuna, salmon, homemade treats and cat food.
An easy recipe is to take 1 pound of ground chicken, 1 cup of bread crumbs (or
flour or cornmeal), 2 Tablespoons of grated cheese, 1 Tablespoon of garlic, mix
together and roll out on a cookie sheet. Bake at 400 degrees for 20 minutes. Let
cool and cut into small pieces. Other meats may be substituted for the chicken
such as a can of tuna or salmon…the consistency of the treats will change
depending on bread crumbs or cornmeal and meats used.
If developing a toy reward system try placing meals in a kong or other toy
device that will encourage the dog to be more interested in the toy. Even if the
dog is already focused on the toy, this training will develop an even greater
interest in the toy. This is a form of classical conditioning used successfully
with frisbee dogs. When using a toy as a reward in training, do not create a
struggle with the dog over releasing the toy. Exchange it with food and make
sure to train the release separately so it is rewarding for the dog to give up
the toy. Many breeds are very possessive over their toy and handlers make this
too big of an issue. The dog often runs away with the toy knowing the handler
will fight for it. Put the toy on a line and pull the dog back in to re-engage
in play. The toy is released and the dog wins but the trainer has the end of the
line and reels the dog back in to play again. When the play session is over and
the trials are finished then allow the dog to go back to the kennel with the
toy.
Be creative with how you reward the animal. Food may be used during the training
session and a jackpot with the toy reward may end a session. The flow of
training must be kept constant as that allows for more behaviors in a timed
trial. More repetition of the target behavior develops the dog’s memory for that
response. If playing with a toy or fighting to get the dog to release a toy
stops the flow in any way, then reward with the toy at the end of the training
session. Once the dog is easily releasing the toy on command then the toy may be
used as a reinforcer during trials and varied with food rewards. Use several
different types of foods and toys so the dog never knows what to expect.
Training is all about keeping the focus of the animal and if the trainer is
unpredictable with the reward system, then the animal is more interested.
Conditioning to people
Next the dog must be conditioned to people. Start by having everyone give treats to the dog at
training. Watch the behavior of the dog and make sure it is comfortable with
people. During the week away from search training the handler should take the
dog to a variety of public locations such as parks, schools, and airports to
develop good social skills with strangers. The handler may ask different people
to give treats to the dog. The dog soon realizes that everyone in its world
loves it and will have good treats to give when it initiates contact. Below is a
list of displacement behaviors to watch for during this phase of training:
- Diverted eye gaze
- Sniffing the ground
- Panting
- Submissive urination or licking
- Pulling away from the stranger
If any of the above behaviors occur, then continue socializing the dog until it
becomes more focused on people. Starting bark alert training at this point would
shut the dog down as it would not be comfortable enough to bark in the face of a
stranger. This would place the dog in conflict and create more displacement
behavior eventually leading to a dog that is unreliable with the bark alert in
the field.
Once the criteria is met with the focus on people giving treats, then have the
dog play toys with people it feels comfortable with. At training sessions the
dog can play with several people by passing off the toy (if tug is the game) to
another person. The handler stands close by but does not engage with the dog at
this time. All play comes from the helpers. The handler is the dog’s security
system so watch to see if the dog acknowledges the handler’s presence and wants
to go back to handler or is able to remain focused on the helpers. Gradually
increase the distance from the handler and dog during the play sessions until
the handler is out of sight and the dog remains engaged with the helper(s). Once
this occurs then the dog is comfortable and is ready to begin the bark alert.
The Bark Alert
The handler begins the shaping of the bark alert. Shaping a behavior means to mark successive approximations to
the final behavior. This means if the dog will only whine, then that is a small
step towards the final behavior, so mark by clicking then reward the whine.
(Remember that the reward must follow the click within a 3 second period of time
as longer delays have been shown to lose the power of the marker signal.)
Eventually the dog will give a stronger response and when that begins to happen
then the handler’s criteria for the bark increases. Next the handler marks only
a bark and not a whine. If at any time the dog’s behavior diminishes, drop
criteria and make it easier to achieve a reward. A training session should only
last for 5 minutes then the dog should be given a break. A young dog may only
have a 3 minute session before resting it due to the shorter attention span.
During the training period the handler should have someone keep track of the
number of responses. This will help determine when to raise or lower the
criteria during the training session.
The quality of the bark should be noted during this time. Make sure to mark only
strong barks…no air barks or whines. The weaker barks will diminish as the dog
understands the type of bark that makes the click occur. Note the frequency of
the bark…are there two or three barks then a pause or is the bark rhythmic? Many
dogs will bounce their paws as they bark. This is the rhythmic bark alert where
the dog responds to a find with a consistent focused bark until the handler
arrives to acknowledge the find. This is the goal to work towards and is
obtainable by doing your ground work and building the foundation of the bark
indication before putting it all together in the search arena.
Once the bark is occurring spontaneously, the handler should increase the
duration and build up to a consistent strong bark that occurs in a rhythmic form
for thirty seconds. During duration building the bark is placed on a variable
reinforcement schedule. At this point the dog should have all strong barks and
the handler is working on marking numbers of barks. For instance the first trial
will be 10 barks and reward, then 5 barks and reward, then 8 barks and reward,
then 20 barks and big reward since the dog worked harder on the last one. The
variable schedule will make the dog work harder as it does not know when the
reward is coming but do not try to push the dog too fast or the behavior will
drop. Read the dog and know when to push and when to drop back. This is similar
to marathon training. One day the run is long and 4 days the run is of shorter
duration. A continual lengthening of duration without dropping back to shorter
numbers will discourage a dog and develop a slower response as the animal knows
it has to do a longer and longer trial before the reward occurs. But an animal
will work harder when there is one long trial and 6 shorter ones with a payoff.
Handlers may be creative with the bark games to play at home. A toy can be
thrown out of reach of the dog and the dog will have to bark to get to it. For
example place food or a toy in the dog’s crate and close the door and let the
dog bark. Reward by opening the door and allow the dog to get to the item of
interest. Have the dog bark for its meal or a treat several times throughout the
day. The handler can run into another room and shut the door and get a bark
response. Use a tennis court where the dog is contained and becomes frustrated
when it can not get to you. Have the dog bark when playing fetch with a toy.
Many repetitions of the bark can be done with simple every day interaction with
the dog. Be careful that you do not reinforce an alarm bark during this
training. This occurs when someone comes to the door and rings the bell. Also
take small approximations to the final behavior as already stated--very
important in shaping. When doing some of these other exercises remember to
initially reward on arrival without a bark, then gradually build up to a bark as
the dog understands the game. Shaping does not mean a bark on the first trial.
In the crate exercise where the ball is thrown into the crate, the first trial
may be a delay in opening for only a couple seconds, then open and let the dog
get to the ball or food without a bark. Do several repetitions of this, then
delay the opening longer to initiate a whine or bark. Do not get the dog too
frustrated in early steps as you may create a fight to get to the ball or begin
a long series of whines before a strong bark. Drop back as needed. When the
response is strong at home the dog is ready to transfer to another person.
Transferring to a Helper
When the dog transfers the bark alert training to another person other than the handler, the dog should
start with someone it knows and is comfortable with. The bark alert will began
from step one as if the dog had never been trained to bark by the handler. The
helper starts with only one bark and builds up to marking strong barks. Do short
five minute sessions and document frequency of barks during 1 minute increments.
A baseline data collection can be done with the handler to determine the number
of barks the handler can get from the dog. The helper should use this criteria
as a goal to work towards. When the helper is getting the same number of barks
from the dog as the handler did in the baseline data then the next step may be
taken.
Use a helper the dog is not as familiar with and again drop back to the
beginning. Note behavior during these training sessions. If any displacement
activity occurs then drop back to just giving the dog treats and develop a
comfort level with the dog before going on to the bark training with the new
person. The handler stands close to the dog during the sessions but does not say
anything. Eventually the handler will increase the distance away from the dog as
the comfort level is built and the dog shows complete engagement with the
helper. This develops confidence in the dog and the ability to work away from
the handler so that handler dependency never becomes an issue.
The helpers are varied before doing different body positions. Once the dog is
relaxed with the exercise then positions and expressions of the helpers are
varied. The helpers will initially look away when trying to get the bark alert,
then they will look directly at the dog. Expressions are changed and unusual
noises are made such as high squeals and moaning. The helper can be dressed up
to look odd with strange hats, glasses, clothing, coats, etc. Positions are
changed to sitting, prone, supine, standing, squatting, and rolling on the
ground and eventually walking. This is desensitization training for anything
unusual the dog will encounter during its search career.
Doing Runaways Short quick runaways (RA’s)
may be done either before the desensitization training or at the same time.
Initially the RA is only several feet. A series of quick drops are done with the
handler running up to retrieve the dog during reward from the helper. The first
series of RA’s is rewarded on arrival to the helper without a bark. The dog
should understand the game after the first trial of quick successive runaways
and immediate reward. Then change to quick successive runaways with three to
five barks on arrival. This is a fun training game for the dog and can be done
as a motivational tool throughout the dog’s training career.
Eventually the RA will lead to an out of sight problem. Remember the new
variable is the RA so lower the criteria for the bark response to only a few
barks. If the bark training has been built correctly, there should not be any
sign of displacement behavior with this step. The criteria is a quick direct
line to the helper without stopping to sniff the ground. If the dog stops to
sniff, then stop the trial immediately without reward. DO NOT allow the dog to
engage in a long repertoire of unwanted behaviors . This is sloppy training and
will carry over in the future. The dog does not have to end on a high note each
and every time. If this becomes the issue then the handler will allow the dog to
engage in this activity which then becomes a long chain of sniffing, messing
around, then bark alert. Be very concise in the final picture and be very clear
to the dog about your expectations. Do Not correct for these behaviors by giving
a verbal “no”, rather take the dog and restart the trial. The correction is not
getting to the toy reward. Ongoing verbal corrections drop the confidence level
of the dog during early phases of training.
Location/Props
Begin hiding the helper in props such as bark barrels, boxes, or under objects such as clothing, pallets,
rubble, wood, etc. Drop back each time a new variable is added. Begin with
runaways into the new objects and leave the doors open. The dog is instantly
rewarded on arriving to the new prop. No bark has to occur until the dog feels
comfortable with the new object. Then slowly build up the bark response at the
prop. This is also how to develop comfort in new training areas. Initially the
dog is introduced to the new object/prop or training area by walking the dog
around the area. Let the dog explore new things first without adding the stress
of barking to the picture. Make sure there are no signs of anxiety with the new
area before beginning the bark training. Always take very small steps when any
variable has changed—topography, new person, or multiple hidden helper problems.
Remember not all dogs are the same. Some will breeze through all the steps
without any issues, some may have a couple issues but will overcome them quickly
and develop a reliable bark alert if shaped correctly and steps repeated when
necessary.
Distraction Training
Distraction training is added slowly to make sure the response of barking has been conditioned in the dog.
Distractions can be in many forms such as other people standing around, dogs
playing nearby, noises such as heavy machinery, animals, and food. Try to keep
control of the situation when adding distractions. If using food then keep it in
a container not accessible to the dog. Start at a distance away from the dog and
make the game of runaway very exciting. Gradually decrease the distance from the
runaway and the distraction until the distraction is only a few feet away from
the helper.
If at any time the dog loses focus on the job at hand then immediately stop the
trial and gain control of the dog. It is not necessary to reprimand the dog. It
is better to show no emotion and collect the dog to restart.
Reassess the situation and determine if the distance from the distraction is too
close. You will determine baseline distance parameters by doing this or in other
words the distance needed before the dog will react to the outside influence.
Keep the distraction at that point or beyond until the dog is repeating the
trials successfully, then decrease the distance and lower the criteria for the
bark. For example increase distance from distraction to the runaway location but
initially reward the dog immediately for coming to the helper. Next trial only
ask for 1 or 2 barks.
One comment on punishment during these trials. Too many dogs completely shut
down and will not even go look for the helper if they are constantly corrected
or punished during distraction training. They may show displacement behavior as
they are placed in conflict. Why does this happen? Well it is due to the dog
investigating a scent and then unsure if they will receive a correction or a
reward. Allow a dog to check out the scent but heavily reinforce when they leave
it. There are also “leave it” training sessions that may be done outside of the
distraction training where the dog is reinforced for leaving anything the
handler asks it to leave. When this type of training is done then the dog may be
told to leave it in an even toned voice—(remember not forceful) during the
distraction training. Unfortunately I have found not many people can remain calm
and non emotional during these sessions, so I do not suggest this for most
handlers.
Keep training exciting and variable. Use several different locations as training
sites so that there is a natural variety of distractions already available. This
prepares the dog for the real world. The final product will be a dog that
maintains its bark alert in all environments with or without distractions. Food
in rubble no longer becomes an issue as the dog understands it receives the food
reward ONLY from the person it is searching for if it is a food reward dog.
Other handlers no longer become concerned when other food reward dogs run before
their toy reward dog. Prepare these search dogs for the real world. There may be
food on any search and any dog will chose to eat it even if it is not a food
reward dog. Dogs are scavengers so do not be fooled into thinking if you never
use food then your dog will never want it! Distraction training will show you
what will happen when you are not around and have no say over what your dog
does. You are teaching the dog how to make the right decision.
This site was last updated on
04/23/2007
|
